Download:
pdf |
pdf
Appendix A
National Survey of Crime Victimization
Companion Study - Pilot
Summary of Pilot Results
Authors
J. Michael Brick
Sharon Lohr
W. Sherman Edwards
Pamela Giambo
Pam Broene
Douglas Williams
Sarah Dipko
March 12, 2013
Prepared for:
Bureau of Justice Statistics
Washington, D.C.
Prepared by:
Westat
1600 Research Boulevard
Rockville, Maryland 20850-3129
(301) 251-1500
Table of Conte
ents
Chapter
1
2
Pa
age
Design
n and Objecttives .........................................................................................
1-1
1.1
1.2
1.3
Pilot Test Sample
S
Design ......................................................................
Pilot Telep
phone Instrum
mentation ........................................................
Additionall Experimentt – Test of Suurvey Name ................................
1-55
1-66
1-88
C
Reesults .......................................................................................
Data Collection
2-1
2.1
2.2
2.3
2.4
2-1
2-55
2-66
2-10
Counts of Yield .......................................................................................
Level of Effort.........................................................................................
Response Rates
R
........................................................................................
Summary of
o Data Colleection Resultts ..................................................
3
Resultts of Other Experiments
E
......................
.
.....................................................
3-1
4
Cost Comparisons
C
s ................................................................................................
4-1
4.1
4.2
4.3
4.4
4.5
5
6
Mail Costs.................................................................................................
Telephonee Costs......................................................................................
Relationsh
hip of Mail an
nd Telephonee Costs ........................................
Effect of Changing
C
Parrameters ............................................................
Compariso
on with RDD
D Costs ......... .....................................................
4-22
4-44
4-88
4-99
4-10
Questtionnaire Asssessment ..................................................................................
5-1
5.1
5.2
5.3
5.4
Overview and
a CATI In
nstrument Tiiming ...........................................
Item Nonrresponse and
d Response F
Frequencies .................................
Intervieweer Debriefingg ..........................................................................
Behavior Coding
C
and Monitoring
M
... .....................................................
5-1
5-33
5-66
5-99
Data Processing
P
.....................................................................................................
6-1
6.1
6.2
6.3
6-1
6-44
6-66
NCVS-CS Pilot
P
Report
Crime Classsification Prrocess ................................................................
Sensitivity Analysis ..................................................................................
Recommen
ndations ..................................................................................
iii
Conten
nts (continued)
Chapter
7
Pa
age
Weigh
hting ................................................................................................................
7-1
7.1
Household
d Weights ................................................................................
7-22
7.1.1
7.1.2
7.1.3
Household
H
Baase Weights ......................................................
Household
H
Nonresponse
N
A
Adjustmentss .............................
Household
H
Po
oststratificatiion ...............................................
7-22
7-33
7-44
Person Weeights ........................................................................................
7-55
7.2.1
7.2.2
7.2.3
Person
P
Base Weights
W
.............................................................
Person
P
Weigh
ht Raking...... .....................................................
Person
P
Weigh
ht Trimming .....................................................
7-66
7-66
7-66
Victimizatiion and Incid
dent Weightss ....................................................
Variance Estimation
E
...............................................................................
7-77
7-77
Estimation Resultss ................................................................................................
8-1
7.2
7.3
7.4
8
8.1
Victimizatiion Rates: Co
omparison w
with Nationall NCVS
Estimates and Estimatees from the C
Chicago Poliice
Departmen
nt ..............................................................................................
8.1.1
8.1.2
8.1.3
8.2
8.3
8.4
8.5
8.6
9
Why
W Might Estimates
E
from
m the NCVSS, the CS,
an
nd the City of
o Chicago bee Expected tto Differ?..............
Property
P
Victiimization ..... .....................................................
Violent
V
Victim
mization ............................................................
Effectiveness of Mail Screener
S
for P
Predicting N
NCVS
Victimizatiion............................................................................................
Nonrespon
nse Bias Asseessment ............................................................
Analysis off Interviewerr Effects ............................................................
Recall Effeects ...........................................................................................
Summary and
a Recomm
mendations ... .....................................................
8-1
8-1
8-44
8-88
8-12
8-15
8-19
8-220
8-221
mary and Nexxt Steps ....................................................................................
Summ
9-1
9.1
9.2
9.3
9-1
9-44
9-66
NCVS-CS Pilot
P
Report
Lessons fro
om the Pilot Study ...............................................................
Objectivess for Further Evaluation . .....................................................
Suggested Approach ...............................................................................
iv
Conten
nts (continued)
Appendices
Pa
age
A
Mail Screeners
S
........................................................................................................
A--1
B
Variattion Between
n Core NCVSS Instrumentts and the W
Westat Pilot ............
B--1
C
Resultts from Behaavior Coding ..........................................................................
C--1
D
Behavvior Coding Questions
Q
Reeviewed Withh Recommenndations ................
D--1
E
Censuus Flow Charrt of TOC Co
odes ...................................................................
E--1
F
Type of
o Crime Speecifications ..............................................................................
F-1
G
Description of thee Type of Criime Codes ........................................................
G--1
H
Censuus Instruction
ns for Using the Automatted NCVS E
Editing and
Codin
ng Blaise Insttrument ....................................................................................
H--1
How To
T Use the Weights
W
....................................................................................
I-11
Comp
parison of Ap
pproach 1 an
nd Approach 2 methodoloogies used
in the Pilot Data Collection
C
.................................................................................
1-22
1-2
Expeccted pilot sam
mple perform
mance, Approoach 1 ..........................................
1-77
1-3
Expeccted pilot sam
mple perform
mance, Approoach 2 ..........................................
1-77
2-1
Numb
ber of sampleed and comp
pleted sampleed addresses,, by
approach..................................................................................................................
2-22
Percen
ntage compleeting household screener of those sennt for
teleph
hone data colllection, by ap
pproach and telephone m
match
status ........................................................................................................................
2-33
Numb
ber of telepho
one numberss sent for teleephone data collection,
by app
proach and phone
p
type ...............................................................................
2-44
2-4
Cases to complete ratio, by app
proach ...............................................................
2-55
2-5
Estimated responsse rates, by ap
pproach ....... .....................................................
2-77
I
Tables
1-1
2-2
2-3
NCVS-CS Pilot
P
Report
v
Conten
nts (continued)
Tables (ccontinued)
2-6
Pa
age
Estimated NCVS-1 household response rattes, by approoach and
characcteristic ...........................................................................................................
2-88
2-7
Estimated Approaach 2 mail ressponse rates, by characterristic ......................
2-99
4-1
Appro
oach 1 mail screener costss .........................................................................
4-22
4-2
Appro
oach 2 mail screener costss .........................................................................
4-44
4-3
Appro
oach 1 teleph
hone and totaal costs ..............................................................
4-55
4-4
Appro
oach 2 teleph
hone and totaal costs ..............................................................
4-77
5-1
Intervviewer and reespondent beehavior codess used in NC
CVS-CS
pilot ..........................................................................................................................
5-10
Generral guide to using
u
the houusehold, person, victimizaation and
incident weights .....................................................................................................
7-77
Estimated propertty victimizatio
on rates, per thousand hoouseholds;
95 perrcent confideence intervalss are given beelow the estiimates.
The laast two colum
mns give the ratios of the CS victimizaation rates
to the NCVS victim
mization ratees from the C
Chicago CBSSA ..........................
8-66
Estimated propertty victimizatio
on rates, per thousand hoouseholds,
for geographic posststrata; 95 percent confiddence intervaals are
given below the esstimates ....................................................................................
8-77
Estimated Violent Victimizatio
on Rates, perr thousand peersons; 95
percen
nt confidencee intervals arre given below
w the estimaates .........................
8-99
Estimated violent victimization
n rates, per thhousand perssons, for
geograaphic poststrrata; 95 perceent confidencce intervals aare given
below
w the estimatees ..............................................................................................
8-99
Estimated violent victimization
n rates, per thhousand perssons, for
demoggraphic domains; 95 perccent confidennce intervals are given
below
w the estimatees ..............................................................................................
8-10
Estimated rates off persons with
h at least onee violent victtimization,
per thousand perso
ons, for geoggraphic postsstrata; 95 perrcent
confid
dence intervaals are given below
b
the esttimates ........................................
8-11
7-1
8-1
8-2
8-3
8-4
8-5
8-6
NCVS-CS Pilot
P
Report
vi
Conten
nts (continued)
Tables (ccontinued)
8-7
Pa
age
Estimated numberr of persons with
w at least one violent
victim
mization, per thousand
t
perrsons, for deemographic ddomains;
95 perrcent confideence intervalss are given beelow the estiimates ...................
8-12
Cross--classification
n of reported
d crimes for hhouseholds oon mail
screen
ner and NCV
VS telephone screener inteerview..........................................
8-13
Sensitiivity and speecificity of the Approach 2 Mail Screener for
NCVSS crimes ..........................................................................................................
8-13
Possib
ble reasons th
hat househollds reportingg crimes on thhe
Appro
oach 2 mail screener did not
n have NC
CVS-2 victimiization
reportts .....................................................................................................................
8-14
Houseeholds reportting victimizaation in the ttelephone intterview
but no
ot in the maill screener.................................................................................
8-14
Comp
parison of NC
CVS-CS Pilo
ot household interview resspondents
with population
p
......................................................................................................
8-16
Comp
parison of NC
CVS-CS Pilo
ot person inteerview responndents
with population
p
......................................................................................................
8-17
Contin
ngency tabless of reportingg at least onee crime on N
NCVS-CS
Pilot telephone
t
quuestionnaire by
b refusal connversion stattus..........................
8-18
Contin
ngency tabless of reportingg at least onee violent crim
me by
refusaal conversion status......................................................................................
8-18
Estimated percentages of victim
mizations repported to pollice, by
o crime...........................................................................................................
type of
8-19
1-1
Flow chart
c
illustratting Approacch 1 Methoddology ...........................................
1-33
1-2
Flow chart
c
illustratting Approacch 2 Methoddology ...........................................
1-44
7-1
Crimee rates from Chicago
C
Police Departmeent ................................................
7-55
8-1
Numb
bers of incideents reported
d by months prior to interrview ....................
8-221
8-8
8-9
8-10
8-11
8-12
8-13
8-14
8-15
8-16
Figures
NCVS-CS Pilot
P
Report
vii
Design
n and O
Objectivves
1
a Key Fin
ndings
Chapter Highlights and
Goa
al to identify a low-cost metthodology to produce
p
local a
area estimatees based on th
he NCVS
insttrument.
Use
e of address-based sampling with telepho
one administrration of the N
NCVS instrume
ents.
Two
o approaches tested in the Pilot study: on
ne using a brieef mail surveyy to obtain tele
ephone numb
bers, a
seccond using a lo
onger mail surrvey to identiffy households at higher risk
k of victimizatiion.
The Bureeau of Justicee Statistics (B
BJS) and Westat have desiigned a Com
mpanion Survey (CS) to th
he
National Crime Victim
mization Survvey (NCVS) as a way of pproducing suub-national estimates usin
ng a
more cost-effective ap
pproach than
n the core NC
CVS. The firrst data collecction phase o
of the NCVSS-CS
was a pilo
ot test in the Chicago-Naperville-Jolieet, IL-IN-WII MSA. The ppilot tested tw
wo data
collection
n approachess, both using an address-b
based samplee (ABS) desiggn. Approach
h 1, a “teleph
hone
number harvest,”
h
screeened by maiil only those addresses foor which we w
were unable to obtain a vvalid
telephonee number fro
om directory services; thee purpose of this mail screeener was prrimarily to ob
btain
a telephone number. Approach
A
2, or the “two--phase ABS hhybrid,” screeened all seleected addresses by
mail with
h a goal of ovversampling households
h
liikely to incluude a victim oof a crime. A
Approach 2 allso
included questions thaat might be used
u
to support model-baased small-arrea estimates (SAE). For both
hes, we develo
oped a teleph
hone version
n of the core NCVS intervview with sam
mpled
approach
househollds, includingg a household
d informant and
a one or tw
wo randomlyy selected aduults. Table 1--1
shows diffferences in design
d
betweeen the Comp
panion Surveey and the coore NCVS, an
nd between
Approach
hes 1 and 2; Figures
F
1-1 and
a 1-2 are fllow chart sum
mmaries of A
Approaches 1 and 2.
NCVS-CS Pilot
P
Report
1-1
Design and
d Objectives
Table 1-1
1.
1
Comparison of App
proach 1 and
d Approach 2 methodolo
ogies used in the Pilot Da
ata
Collecction
Desig
gn Elements
Design
Scope
Frame
Clustering
Household (HH)
subsamp
pling
Within-HH
H selection
Co
ompanion Survvey
A
Approach 1
Approach 2
NCVS
Sample
Panel
Cross-ssection
National
MSA(s )
Area
ABS
Multi-stag
ge geographicc
None
None
None
Matche
ed addresses not
respon
nding to mail @ 50%
H; No adolesce
ents
Up to 2 adults in HH
Mode
All adultss in HH; All
adolescents aged 12-1
17
Da
ata Collection
In-person
n/telephone
Mail/teelephone
Interview
wers
Respondent incentive
Census FRs
None
Mail scre
eener
Household
enumera
ation
Victimiza
ation screener
Victimiza
ation details
None/W
Westat teleph
hone
None
Instruments
I
None
Unmattched addressses
All add
dresses; Appro
oach 1
only; N
Neighborhood
plus vicctimization an
nd
questioons plus
correla
ates
telephoone number
Control Card,
C
householld
Adapteed for telepho
one
informant
NCVS-1, household
h
Adapteed for telepho
one; some item
ms dropped
informant
NCVS-1, other
o
sampled
d
Adapteed for telepho
one
person(s))
NCVS-2 (IIncident Repo
ort)
Adapteed for telepho
one
NCVS-CS Pilot
P
Report
Telephone only for
matche
ed addresses
1-2
Design and
d Objectives
Figure 1-1
1.
Flow chart illustratting Approacch 1 Methodoology
NCVS-CS Pilot
P
Report
1-3
1
Design and
d Objectives
Figure 1-2
2.
Flow chart
c
illustratting Approacch 2 Methodoology
ABS Frame
ABS Sample
Mail screener to
M
address
Sccreener returned
with
h phone number?
?
No
Match to get
g phone
number
Matching
g phone
numb
ber?
Yes
Yess
Classify HH
Yes
Screener re
eturned?
No
o
High Risk
Low Risk
Sample at
a Rate
TBD
D
Send to TRC for HH
interview.
NCVS-CS Pilot
P
Report
Sample ½ of
S
HHs
1-4
No
Nonre
esponding HH
1
Design and
d Objectives
1
The goalss of the pilott were to:
1.1
Assess th
he viability off ABS is obtaaining sub-naational estimaates in a costt-effective
manner.
Identify whether
w
App
proach 1 or Approach
A
2 pprovides morre informatio
on for produccing
blended estimates.
e
Identify which
w
approaach provides more inform
mation for sm
mall area estim
mation.
Analyze the
t effectiven
ness of the Approach
A
2 sccreener at identifying houuseholds with
ha
victim.
ne optimal suubsampling fractions
fr
for a subsequentt full-scale teest.
Determin
Pilot Test
T
Sample Desig
gn
The desiggn for both approaches
a
sttarted with a stratified sim
mple random
m sample of aaddresses seleected
from the ABS frame in
i the Chicaggo-Napervillee-Joliet, IL-IN
N-WI MSA. The ABS fraame is a file o
of
residentiaal addresses that
t is maintaained by a veendor, based on the Uniteed States Posstal Service
(USPS) Computerized
C
d Delivery Seequence File (CDSF). Tw
wo strata weree created, on
ne for the cen
ntral
city of Ch
hicago and on
ne for the rest of the MSA. The samppling rate in tthe central ciity was 1.5 tim
mes
the rate in
n the remainder of the MSA
M to increaase the expeccted yield of vvictims of crrime and imp
prove
the abilityy to make comparisons with
w the Coree NCVS. Thee sample was randomizedd within stratuum
to the two
o principal exxperimental conditions.
oach 1, addreesses were sam
mpled from the ABS fram
me and immeediately matcched by a ven
ndor
In Appro
to identiffy telephone numbers
n
associated with the addressees. Those witth matching telephone
numbers were sent an
n advance lettter, and sent for a telephoone interview
w. If no telep
phone numbeer
was availaable for a sam
mpled addresss, then a maail screener w
was sent to thhe address. Iff a matched
telephonee number pro
oved to be out of service or not assocciated with thhe sampled aaddress, we allso
sent a maail screener to
o the addresss. The primarry purpose oof the mail screener in this approach w
was
to obtain a telephone number. To increase inteerest in the s urvey, we inccluded a limiited number of
questionss on perceptions of the neeighborhood
d and of emergency servicces. Those ho
ouseholds th
hat
respondeed and provid
ded a telepho
one number were
w sent (baack) for telepphone interviiewing.
In Appro
oach 2, all add
dresses samp
pled from thee ABS frame were also m
matched with telephone
directoriees. However, both matcheed and unmaatched addresses were sennt a mail screeener. The
NCVS-CS Pilot
P
Report
1-5
Design and
d Objectives
1
screener included
i
the same questio
ons as the Ap
pproach 1 sccreener, as weell as a few qquestions on
victimizattion experien
nces and charracteristics asssociated witth a higher likkelihood of vvictimization
n.
These items were used
d to classify households
h
as
a either Higgh Risk (likelyy to have exp
perienced
victimizattion in the paast year) or Low
L Risk (un
nlikely to havve experienceed victimizatiion). The inteent
was to sample househ
holds classifieed as High Risk
R with certtainty and as Low Risk at a rate of ½.
Those sellected would
d be sent for the
t telephon
ne interview. Non-responnding households with
matched telephone nuumbers weree also to be saampled at ann initial rate oof ½ for telep
phone
interview
wing.
mples were released in rep
plicates for mailing
m
and foor telephone interviewingg if appropriaate.
Both sam
The purp
pose of the reeplicates was to allow adjuustments to tthe subsamppling rates if n
needed to achieve
target num
mbers of com
mpleted interrviews. In practice, lowerr-than-expectted response to the mail
survey resulted in abaandoning the subsamplingg by risk cateegory.
As part of
o the telepho
one interview
w, two adults were random
mly sampled from househ
holds with th
hree
or more adult
a
residents. If there were
w only onee or two adullts, they weree sampled wiith certainty.
Pilot sam
mple sizes and
d expected saample perform
mance for thhe two approoaches are sum
mmarized in
n
Tables 1--1 and 1-2.
1.2
Pilot Telephone
T
e Instrum
mentation
n
The telep
phone interviiew includes three separatte instrumennts: a househoold screener, a victimizatiion
screener asked
a
of each
h sampled peerson, and an
n incident repport asked abbout each rep
ported
victimizattion. These correspond
c
to
o the core NCVS
N
Controol Card, NCV
VS-1, and NC
CVS-2,
respectiveely.
The houssehold screen
ner includes a household roster and deemographic questions fro
om the NCV
VS
Control Card,
C
and staandard questiions to identiify whether aan eligible hoousehold at th
he sampled
address had
h been reacched. It is con
nducted with
h an adult (188 or older) livving in the h
household. Att the
end of th
he household screener, thee sampled ad
dults were askked to responnd to the vicctimization
screener (NCVS-1, minus
m
question
ns on identityy theft). Thee household sscreener resp
pondent was
asked to complete
c
both the household and perrsonal victim
mization questtions, whether or not s/h
he
was samp
pled. Any oth
her sampled adult
a
was askked to compllete only the section of th
he victimizatiion
NCVS-CS Pilot
P
Report
1-6
Design and
d Objectives
1
screener covering
c
personal crimes. Following completion
c
oof each victim
mization screeener, the
respondeent was asked
d to completee an incident report for eaach victimizaation mentio
oned. The inccident
report rep
plicates the NCVS-2
N
as cllosely as possible in the telephone envvironment.
Table 1-2
2.
Expected pilot sam
mple perform
mance, Approoach 1
Sample components
c
Addressess sampled
Vacancy Rate
R
Occupied households
Vendor ph
hone number match
m
rate
Household
ds matched fo
or phone #
% of Vendor phone num
mbers that are valid
Household
d screeners mailed
m
Household
ds where interrview attempted
Household
d interview ressponse rate
Expected household
h
interview comple
etes
Avg. # of adults
a
sample
ed per househo
old
Extended person interviiew response rate
Expected extended
e
persson completess
Table 1-3
3.
Asssumptions
1
12,500
12%
1
11,000
60%
6,600
80%
5,720
6,996
35%
2,449
1.706
75%
3,134
Expected pilot sam
mple perform
mance, Approoach 2
Sample components
c
Addressess sampled
Vacancy Rate
R
Occupied households
% of Houssehold screene
ers returned with
w phone num
mber
Household
d screeners re
eturned with phone numberr
% of house
ehold screene
ers returned with
w no phone number
Household
d screener returned with no
o phone number
Vendor ph
hone number match
m
rate
Household
d screeners co
ompleted usin
ng vendor phon
ne number
Subsampling rate for sccreener nonresspondents
Household
d screener non
nrespondents subsampled
Household
d screeners co
ompleted in High Risk Strattum (25%)
Household
d screeners co
ompleted in Lo
ow Risk Stratu
um (75%)
Subsampling rate for High Risk Stratum
Subsampling rate for Lo
ow Risk Stratu
um
Household
ds subsampled in Low Risk Stratum
Household
ds sent for tele
ephone intervview
Household
d interview ressponse rate
Expected household
h
interview comple
etes
Average #adults
#
sample
ed per househ
hold
Extended person interviiew response rate
Expected extended
e
persson interview completes
c
NCVS-CS Pilot
P
Report
1-7
Asssumptions
1
14,000
12%
1
12,320
40%
4,928
5%
616
50%
308
1 in 2
1,694
1,309
3,927
1
1 in 2
1,964
4,967
50%
2,483
1.706
75%
3,179
Design and
d Objectives
1.3
1
Additio
onal Expe
eriment – Test off Survey N
Name
In additio
on to testing the two diffeerent samplin
ng approachees, we implem
mented a test of the
Approach
h 2 mail screener to assesss whether “localizing” thhe instruments would havve any effect on
response rates and ressponse patterrns. Half of the
t sample alllocated to A
Approach 2 w
was randomlyy
assigned to
t the localizzed version (w
which uses th
he term “Chiicagoland” inn the survey name and
1
FAQs ) and
a the remaiinder receiveed a generic version
v
of thiis instrumennt2.
1
The term “C
Chicagoland” was received
r
favorably by
b cognitive interviiew subjects in both
th the city of Chicag
ago and outlying sub
burbs.
2
Since the Ap
pproach 1 instrumeent was only mailed
d to those without a matching telephoone number only oone version (local) o
of this instrument w
was
used.
NCVS-CS Pilot
P
Report
1-8
Data
D
Co
ollectio
on Resu
ults
2
a Key Fin
ndings
Chapter Highlights and
ABS
S sample wass matched to telephone
t
num
mbers using d irectory vendo
ors. The match
h rate was low
wer
tha
an initially projjected. In addiition, fewer re
espondents to the mail survvey provided a telephone nu
umber
tha
an anticipated.
Ressponse to the mail survey in
n Approach 2 was reasonab
ble and could be further imp
proved with
add
ditional mailin
ngs and a sma
all incentive.
Hou
usehold level response to th
he telephone interview wass low, with a siizeable numbe
er of househo
olds
nevver answering the telephone
e. Once a houssehold compl eted the screeener, the persson-level respo
onse
wass high for the screener resp
pondent, but was
w low for oth
her sampled a
adults.
The
e tested appro
oaches were essentially equ
uivalent in oveerall weighted response rate
e.
This sectiion presents the results of
o the pilot daata collectionn yield (numbber of complleted intervieews)
and respo
onse rates. Beecause the piilot was desiggned to test ttwo differentt approachess, all of the reesults
are given separately fo
or Approach 1 and Appro
oach 2. In thhe final sectioon, we presen
nt the results of
the secon
ndary experim
ments.
2.1
Countss of Yield
d
Table 2-11 shows the number
n
of saampled addreesses and thee outcomes fr
from each staage of data
collection
n by approach
h. As noted earlier,
e
Apprroach 2 had a larger sampple because o
only subsamp
ples
of (1) tho
ose identified
d as having lo
ow risk of rep
porting a victtimization annd (2) nonresspondents to the
mail survvey with a ven
ndor-matched telephone number werre to be follow
wed up by teelephone. Th
he
former suubsampling was
w not carrieed out, so alll returned maail screeners in Approach
h 2 with a ph
hone
number were
w sent forr telephone data
d collection
n.
Table 2-11 shows that only about 40
4 percent off addresses w
were matchedd to telephon
ne numbers,
which waas lower than
n expected; th
he fact that th
he rate in Chhicago was m
much lower m
may be the no
orm
for large metropolitan
n areas. The percentage
p
of returned m
mail screenerss differs by ap
pproach becaause
the samplles differed in the mix off matched and
d unmatchedd addresses ((discussed beelow). About 74
NCVS-CS Pilot
P
Report
2-1
Data Collecttion Results
2
percent of
o the returneed mail screeners included
d a valid 10-ddigit telephonne number. T
This rate wass also
lower thaan we expecteed; the differrence may aggain be due too the locationn of the sample.
o Table 2-1 shows that despite
d
the laarger sample,, fewer Apprroach 2 cases were
The next--to-last row of
sent for telephone
t
datta collection because of th
he 50 percennt subsamplinng of mail no
onresponden
nts
with a maatched numb
ber. Finally, a higher perceentage of Appproach 2 casses than App
proach 1 casees
was comp
pleted by teleephone; again
n, this differeence is relateed to the subssampling.
Table 2-1
1.
Numb
ber of sample
ed and comp
pleted sampl ed addresse
es, by approach
Addressess sampled
Telephone
e numbers ma
atched
Sent a ma
ail screener
Completed
d mail screene
er
Telephone
e number repo
orted in screen
ner
Sent for te
elephone data
a collection
Completed
d household telephone scre
eener
App roach 1
Number
Percent
12,500
5,142
41%
7,638a
61%a
2,518
33%
1,847b
73%
6,909
1,261
18%
A
Approach 2
Numbe
er
Perce
ent
14,00
00
5,57
72
40
0%
14,00
00
5,41
19
39
9%
4,00
04
74
4%b
6,01
12c
1,35
50
22
2%
Source: 2012 NCVS Comp
panion Survey Pilot
P
a
Includes vendor
v
telephon
ne matches tha
at were not in se
ervice or for thee sampled addrress
b
Excludes telephone
t
numbers obtained too
t late to mak
ke calls
c
Includes subsample
s
of mail
m nonresponsse with vendor telephone
t
num bers
To comp
pare the Apprroach 1 and Approach
A
2 mail
m responsse, we restrictt the sample to those cases
without a matched tellephone num
mber, becausee these cases should havee the same response
propensitty. There werre 7,358 unm
matched Apprroach 1 addrresses sent thhe shorter Ap
pproach 1 maail
screener. Of these, 333 percent werre completed
d. For Approoach 2, there were 8,428 uunmatched
addressess, of which 33 percent weere completed
d using the loonger mail sccreener. Thuus, the two
approach
hes were equaally effective in obtaining mail responnses for unmaatched addreesses.
The overall Approach
h 2 mail completion rate in
i Table 2-1 is higher thaan the Appro
oach 1 rate
because many
m
fewer Approach
A
1 matched
m
addresses were ssent the screeener. The ratte for matcheed
addressess in Approach
h 2 was 47 percent
p
(2,6300 of 5,572). T
The rate for m
matched adddresses in
Approach
h 1, those with “bad” teleephone numb
bers, was 20 percent (57 of 280). (Theese numbers are
not show
wn in the table.)
Table 2-22 presents mo
ore detail on results from
m the telephonne follow-upp, separately b
by approach and
by wheth
her there was a matched teelephone num
mber. The laast row for eaach approach
h shows the
NCVS-CS Pilot
P
Report
2-2
Data Collecttion Results
2
percentagge of sampled
d addresses that
t completeed the houseehold screeneer, where thee denominato
or is
the numb
ber of cases sent
s for telep
phone follow-up. Comparring the resullts by approaach, we see liittle
differencee for unmatcched addressees (34 percen
nt Approach 1 versus 33 percent App
proach 2). Th
here
is a differrence for the matched casses (13 perceent Approachh 1 versus 177 percent App
proach 2), wh
hich
is entirelyy due to subssampling App
proach 2 maiil screener noonrespondennts.
Comparin
ng the resultss by match sttatus, the com
mpletion ratee was twice aas high for un
nmatched
addressess as for match
hed addressees. However, the unmatchhed cases aree only those rreturning a m
mail
screener with
w a teleph
hone numberr, so are the most
m cooperaative part of the overall uunmatched
sample. We
W can comp
pare househo
old screener completion
c
rrates for matcched mail resspondents
providingg a telephonee number (288 percent, datta not shownn in Table 2-22) with unmaatched mail
respondeents (33 perceent) in Appro
oach 2. This result is som
mewhat surprrising, since m
matched casees
typically are
a more coo
operative than
n unmatched
d cases.
Table 2-2
2.
Percentage complleting househ
hold screeneer of those se
ent for teleph
hone data
collecttion, by apprroach and tellephone mattch status
Approach 1
Sent for te
elephone data
a collection
With co
ompleted mail screener
Completed
d household screener
s
Numbe
er
Percen
ntage
Approach 2
Sent for te
elephone data
a collection
With co
ompleted mail screener
Completed
d household screener
s
Numbe
er
Percen
ntage
Total
Ma
atched
Unmatched
d
6,909
1,824
5,,142
57a
1,767
1,767
668
1
13%
593
34%
3,,970
2,,593
2,042
2,042
684
1
17%
666
33%
1,261
18%
%
6,012
4,635
1,350
22%
%
Source: 2012 NCVS Comp
panion Survey Pilot
P
a
Includes 12
1 completed screeners
s
that did
d not provide a valid phone n
number.
As mentioned above, in Approach
h 2 nonrespo
ondents to thhe mail surveyy with match
hed telephon
ne
numbers were subsam
mpled for teleephone follow
w-up. This pprocess was nnot very succcessful, with only
83 of the 1,413 subsam
mpled nonreespondents completing thhe householdd telephone sscreener
interview
w. The latter number
n
incluudes 63 addreesses for whiich a mail scrreener was reeturned after the
subsamplling had occuurred, of who
om 5 compleeted the houssehold screenner on the teelephone.
NCVS-CS Pilot
P
Report
2-3
Data Collecttion Results
2
The mostt substantial difference beetween the tw
wo approachhes was the tyype of phonee number
available for data colleection. Tablee 2-3 shows the
t number aand proportioon of telepho
one numberss
used in caalling the houuseholds by approach
a
and
d phone typee (landline orr cell). The ro
ows show
whether telephone
t
nuumbers were obtained fro
om the vendoor match, thee mail surveyy, or both. In
n
Approach
h 1, the mail survey was sent
s only if th
here was no m
matched num
mber or the m
matched num
mber
was nonw
working or no
ot associated
d with the sam
mpled addresss. In Approach 2, when a mail surveyy was
returned with a teleph
hone numberr, that numbeer was used ffor the telephhone follow--up regardless of
whether or
o not there was
w also a matched
m
numb
ber. So, for tthe “both” roows, the num
mbers and
proportio
ons representt the telephone type of th
he number obbtained in thhe mail.
Only 1 peercent of num
mbers obtain
ned by match
hing were forr cell phones,, because theere is no
equivalen
nt of the Whiite Pages avaiilable for celll phone num
mbers. The sm
mall percentaage of matcheed
cell phon
ne numbers may
m have beeen landline nuumbers that hhouseholds pported to celllular service..
Consequeently, Approach 1 had muuch lower peercentage of cell numberss than Appro
oach 2 of celll
phone nuumbers sent for
f data colleection as App
proach 2 (16%
% and 27%, respectively)). In terms
completeed household
d screeners, Approach
A
1 had
h 33 percennt done on ccell phones an
nd Approach
h2
44 percen
nt.
Table 2-3
3.
Numb
ber of telepho
one numberss sent for teleephone data
a collection, b
by approach and
phone
e type
Total
Only num
mber from ma
atch
Only num
mber from ma
ail
Number from match and
a mail
Percentag
ge cell phoness
Only num
mber from ma
atch
Only num
mber from ma
ail
Number from match and
a mail
Approach 1
Land
dline
Cell
5,7
780
1,129
5,0
062
35
673
6
1,094
45
0
16%
1%
62%
0%
App
proach 2
Landline
Cell
4,407
1,616
6
2,002
15
5
843
1,199
9
1,562
402
2
27
7%
1
1%
59
9%
20
0%
Source: 2012 NCVS Comp
panion Survey Pilot
P
Another measure
m
of yield
y
or produuctivity is a statistic
s
calledd the “cases tto complete ratio,” which
h is
defined as
a the numbeer of sampled
d cases divideed by the num
mber of com
mpleted cases. The lower tthe
ratio, the more efficieent the data collection
c
sch
heme. Table 22-4 shows thhe cases to co
omplete ratio
o by
h, where a com
mplete is deffined differen
ntly for each row. Even tthough we suubsampled m
mail
approach
nonrespo
ondents with matched teleephone numbers rather thhan sending all of them ffor telephonee
follow-up
p, the numeraator of the caases to comp
plete is all sam
mpled cases, not just thosse subsampleed.
NCVS-CS Pilot
P
Report
2-4
Data Collecttion Results
Table 2-4
4.
2
Casess to complete
e ratio, by approach
A
Approach 1
10
11
46
417
Household
d screener (NC
CVS Control Ca
ard items)
Victimization screener (NCVS-1)
(
Victimization screener with
w reported property crime
Victimization screener with
w reported personal crim
me
Approach 2
10
0
10
0
48
8
609
9
Source: 2012 NCVS Comp
panion Survey Pilot
P
The first row in Tablee 2-4 shows that
t the num
mber of samplled addressess needed to ccomplete onee
householld screener was
w about 10 for each app
proach. The ssecond row ddisplays the ccases to comp
plete
for at leasst one victim
mization screeener in a houusehold, whicch is virtuallyy the same as the ratio forr the
control caard. Nearly all
a household
ds that complleted the houusehold screeener were willling to comp
plete
at least on
ne victimizattion screener. (About 93 percent
p
of hoousehold scrreener respon
ndents went on to
completee a victimizatiion screener..)
The next two rows sh
how the ratios where the complete
c
is ddefined as at least one rep
ported propeerty
crime or at least one reported
r
perssonal crime in the househhold. These aare crimes th
hat wound up
p
being classsified as pro
operty or personal crimes after codingg. (See Sectioon 6.) Because of the relattive
rarity of personal
p
crim
me reports, th
he ratios are much
m
higherr for this typee of complette. With the C
CS
pilot dataa collection procedure,
p
a sample
s
of about 400 to 6 00 cases is nneeded to pro
oduce a singlee
householld with a perssonal crime report.
r
2.2
Level of
o Effort
As we no
oted previoussly, the mail data
d collectio
on involved ssending a pacckage with th
he screener an
nd
no monettary incentivee for both ap
pproaches. (IIn Approach 1, if a matchhed telephon
ne number waas
available only an advaance letter waas sent to thee address beffore calling itt.) About 10 days later, a
thank youu/reminder postcard
p
wass mailed to alll addresses thhat were maiiled the mail screener. Ab
bout
2 weeks later, a secon
nd package was sent to thee nonresponnding addressses, again witthout any
incentive.
nding to a hiigh
For the teelephone datta collection, the level of effort was reelatively intennse correspon
effort ran
ndom digit diial survey. On
n average, a little
l
more thhan 9 call atteempts were m
made per
completeed household
d screener forr both approaches. This nnumber incluudes calls maade to all
househollds sent for teelephone folllow-up. It do
oes not include additional
al calls made tto attempt
NCVS-CS Pilot
P
Report
2-5
Data Collecttion Results
2
completio
on of victimiization screen
ners and inciident reports. Many of thhe interviews were finalizeed
with a sm
mall number of
o attempts, especially
e
tho
ose cases thaat were eventtually compleeted and thosse
with non--working num
mbers. When
n a household refused, a rrefusal conveersion was atttempted afteer a
week or two
t unless th
he first refusaal was classifiied as hostilee. If the houssehold refuseed a second ttime
it was nott attempted again.
a
Aboutt 45 percent of
o all cases atttempted in bboth approacches refused at
least oncee. Just under a third of alll completed household
h
sccreeners were the result o
of a refusal
conversio
on attempt.
Once thee sampled aduult(s) was/weere identified
d, we attemptted to compllete a victimiization screen
ner
with the household
h
sccreener respo
ondent and any (other) saampled adult(s). Even wh
hen the houseehold
screener respondent
r
was
w not samp
pled, we askeed the househhold screeneer respondentt to competee a
victimizattion screenerr; the data weere used for household
h
(pproperty) crim
me purposess. No data fro
om
non-samp
pled adults were
w used forr personal criminal victim
mizations. Thee household respondent
completeed the victimiization screen
ner about 933 percent of tthe time. Thee completion
n rate for aduults
other than the househ
hold respond
dent was mucch lower, rannging from 200 percent to 33 percent
dependin
ng on the num
mber of interrviews attemp
pted in the hhousehold. Thhis result, muuch lower
completio
on rates when having to go
g to a differrent adult witthin the houssehold, is verry common iin
telephonee interviews.
2.3
Respo
onse Rate
es
In this section we preesent responsse rates by ap
pproach for tthe mail (onlyy Approach 22) and teleph
hone
A of the resp
ponse rates are
a weighted by the base w
weights. (Seee Section 7.) F
For Approacch 1
efforts. All
the base weight
w
for co
omputing thee telephone NCVS-1
N
respponse rate is the inverse o
of the probab
bility
of selectio
on for the ad
ddress; this iss also the basse weight forr the mail datta collection effort for
Approach
h 2. For the Approach
A
2 telephone
t
response rates , the base weeight is the product of thee
inverse of the probability of selecttion of the ad
ddress and thhe inverse off the subsamp
pling rate forr
those thaat were subjecct to being suubsampled.
Some of the sample addresses
a
werre no longer associated w
with occupiedd householdss, so these cases
are not bee counted in the denomin
nator of the response
r
ratee computatioons. When th
he householdds
were mailled the instruument, the po
ost office retturned some of these as nnon-deliverab
ble. Using th
he
non-delivverable return
ns (along witth some of th
he telephone dispositionss for Approacch 1 cases), w
we
estimated
d that about 13
1 percent off all the samp
pled addressees were for aaddresses thaat were not
NCVS-CS Pilot
P
Report
2-6
Data Collecttion Results
2
eligible fo
or the surveyy (primarily vacant units). Thus, we muultiplied the weights of th
he
nonrespo
ondents (onlyy those that never
n
returneed the mail) bby 0.87 in acccordance witth AAPOR’s
response rate guidancce to computte what they refer
r
to as RR
R3 (AAPOR
R 2011).
We treat the responsee rates as rand
dom variablees because w
we are interestted in compaaring the ratees
from onee approach to
o the other, and
a we need to understannd the variation that migh
ht be expecteed if
these app
proaches were implementted in the future. Thus, w
we present sam
mpling errors for the resp
ponse
rate estim
mates.
Table 2-55 shows that the responsee rate for the mail screeneer for Approoach 2 was 466.5 percent, w
which
is relativeely high given
n no monetarry incentive and
a only twoo questionnaiire mailings. (A postcard
reminder was also sen
nt to all households via Fiirst Class Maail.) Monetaryy incentive, aadditional
mailings, and use of special mail arre methods that
t tend to iincrease the m
mail responsse rate. No m
mail
response rate is given for Approacch 1; becausee only those w
without a venndor match or with a “baad”
matched number weree included, comparing th
he approache s is not apprropriate.
Table 2-5
5.
Estima
ated responsse rates, by approach
a
Mail scree
ener response rate
Telephone
e response rattes
Househo
old screener ra
ate
Conditional victimization screener rate
r
Approacch 1
Estimate
S.E.
–
12.3%
68.0%
0.3%
–
Ap
pproach 2
Estimate
S.E.
46.5%
0.5
5%
12.4%
64.3%
0.4
4%
–
Source: 2012 NCVS Comp
panion Survey Pilot
P
The houssehold screen
ner telephonee response raate is just oveer 12 percentt for each approach. Thiss rate
indicates that an adultt in the houseehold complleted the houusehold screeener interview
w. There is
nce in this keyy rate betweeen the two appproaches.
essentiallyy no differen
The last row
r of Tablee 2-5 shows the condition
nal person-levvel response rate; this rate is the weigh
hted
percentagge of all samp
pled adults who
w completeed the victim
mization screeener. It is con
nditional beccause
the denom
minator is th
he (weighted) number of sampled
s
adullts and does not account for earlier sttages
of nonressponse. As iss typical of most
m telephon
ne surveys, thhe person completing thee household
screener was
w very likeely to continuue and compllete the victim
mization screeener. The co
onditional
completio
on rate for th
hese adults was
w over 90 percent
p
regarddless of how
w many adultss were samplled
for intervviews. As a reesult, propertty crimes couuld be estimaated from nearly all the households th
hat
NCVS-CS Pilot
P
Report
2-7
Data Collecttion Results
2
completeed the househ
hold screenerr, and the ressponse rate ffor estimatingg property crrimes is essen
ntially
the same as the houseehold screeneer response rate.
r
The cond
ditional respo
onse rate for other adults,, i.e., an adultt other than the one com
mpleting the
householld screener, was
w only arouund 30 percent, and whenn 3 adults weere scheduledd for intervieews
the comp
pletion rate was
w less than 20 percent. Households
H
w
with 3 or moore adults aree more likely to
include yo
oung adults still
s living with their paren
nts or in rooommate situaations, and th
hese adults tend to
be at a higher risk of being
b
victimiized than old
der adults. Inn fact, the perrcentage of aadults reportiing
being a viictim was higgher when 3 adults were scheduled
s
foor interview aand those intterviews weree
completeed.
Table 2-66 presents thee estimated response
r
ratees to the household screenner interview
w by
characteristics of the sampled
s
casees and by app
proach. Sincee the approacch was the m
main experim
mental
condition
n, it is worth noting that overall
o
and accross all the characteristiccs, the response rates forr the
two appro
oaches are nearly identicaal and there is
i no indicatiion of one appproach givin
ng a higher
response rate than thee other by an
ny of these ch
haracteristicss.
Table 2-6
6.
Estima
ated NCVS-1
1 household response
r
ratees, by approa
ach and characteristic
Overall
Vendor ph
hone match
With ven
ndor matched phone
Without vendor match
hed phone
Areas by Census
C
2010 Tract
T
Tract witth 30% or morre Blacks
Tract witth less than 30
0% Blacks
Tract witth 30% or morre Hispanics
Tract witth less than 30
0% Hispanics
Tract witth 30% or morre Black and Hispanic
H
Tract witth less than 30
0% Black and Hispanic
Areas by crime
c
rates
Chicago low crime are
ea
Chicago medium crim
me area
Chicago high crime arrea
Remaind
der of Cook co
ounty
CBSA ou
utside of Cook
k county
Ap
pproach 1
Estimatee
S.E.
12.3%
%
0.3%
%
Source: 2012 NCVS Comp
panion Survey Pilot
P
NCVS-CS Pilot
P
Report
2-8
A
Approach 2
Estima
ate
S.E
E.
12.4
4%
0.4
4%
14.7%
%
10.2%
%
0.4%
%
0.4%
%
15.2
2%
10.1
1%
0.6
6%
0.4
4%
10.6%
%
12.6%
%
11.7%
%
12.4%
%
11.4%
%
12.8%
%
0.7%
%
0.3%
%
0.7%
%
0.3%
%
0.5%
%
0.4%
%
10.6
6%
12.6
6%
11.7
7%
12.4
4%
11.4
4%
12.9
9%
0.7
7%
0.3
3%
0.7
7%
0.3
3%
0.5
5%
0.4
4%
11.6%
%
11.0%
%
9.1%
%
12.8%
%
13.1%
%
0.7%
%
0.9%
%
0.7%
%
0.6%
%
0.5%
%
11.6
6%
11.0
0%
9.1
1%
12.8
8%
13.1
1%
0.7
7%
0.9
9%
0.7
7%
0.6
6%
0.5
5%
Data Collecttion Results
2
If we com
mpare rates by
b characterisstics of the ad
ddresses withhin approachh, there is som
me variation.. One
of the mo
ost important factors thatt consistentlyy influences rresponse proopensities in m
most mail an
nd
telephonee surveys is the ability to match the saample addresss to a telephhone number (for telepho
one
samples the
t equivalen
nt is matchingg to an addreess). In both approaches, the addressees that could be
matched to a telephon
ne number had
h response rates that weere about 5 ppercentage po
oints higher3. The
next set of
o response rates
r
is based
d on the perceentage of thee population who are Blaack and/or
Hispanic.. The areas arre classified using
u
data fro
om the 20100 Census tracct of the samp
pled addresses.
While thee high Black and the high
h combined Black
B
and Hi spanic areas show significantly lower rates
than the other
o
areas, the
t differences are all sub
bstantively sm
mall.
The last section
s
of thee table gives the responsee rates by crim
me rates repoorted by the Chicago pollice
departmeent. (The areaas outside off the city are not
n classifiedd by crime raates.) These aare the catego
ories
used in th
he poststratiffication step of weightingg, and are disccussed in moore detail in SSection 7. Th
he
general pattern is as expected, with
h lower respo
onse rates inn higher crime areas, but tthe differences
are relativvely small and
d only a few of the differrences are staatistically signnificant for b
both approacches.
Table 2-7
7.
Estima
ated Approach 2 mail ressponse ratess, by characte
eristic
Overall
Vendor ph
hone match
With ven
ndor matched phone
Without vendor match
hed phone
Areas by Census
C
2010 Tract
T
Tract witth 30% or morre Blacks
Tract witth less than 30
0% Blacks
Tract witth 30% or morre Hispanics
Tract witth less than 30
0% Hispanics
Tract witth 30% or morre Black and Hispanic
H
Tract witth less than 30
0% Black and Hispanic
Areas by crime
c
rates
Chicago low crime are
ea
Chicago medium crim
me area
Chicago high crime arrea
Remaind
der of Cook co
ounty
CBSA ou
utside of Cook
k county
Estimatee
46.5%
%
S.E.
0.5%
53.0%
%
41.2%
%
0.7%
0.7%
37.6%
%
48.2%
%
38.5%
%
48.3%
%
38.9%
%
51.4%
%
1.1%
0.5%
1.0%
0.6%
0.7%
0.7%
44.8%
%
39.4%
%
36.4%
%
49.2%
%
48.7%
%
1.0%
1.2%
1.2%
1.0%
0.7%
Source: 2012 NCVS Comp
panion Survey Pilot
P
3
Unless otherwise noted, wheneever we describe a difference as beingg higher or lower inn this report, it meaans the difference iis statistically signifficant
using a two--sided t-test with =0.05. No adjustm
ments have been maade for multiple tessts.
NCVS-CS Pilot
P
Report
2-9
Data Collecttion Results
2
Table 2-77 gives the mail
m response rate for App
proach 2 by thhe same charracteristics ap
pplied to
telephonee response raates in Table 2-6. The pattterns noted for the household screen
ner are even m
more
apparent for the mail response rattes. For exam
mple, the addrresses with m
matched telep
phone numb
bers
r
ratee than those w
without a maatch. While aat
have abouut a 12 perceentage point higher mail response
first this may
m suggest a greater pottential for biaas for the maail survey, in relative term
ms this is not the
case. Thee mail rates have
h
a ratio of 1.3 (53.0/441.2=1.3) whhile the telephhone rates haave a ratio off 1.5
(15.2/10.1=1.5, for Approach
A
2). The ratio is more
m
indicattive of the pootential for n
nonresponse b
bias
than is th
he absolute diifference in rates.
r
2.4
Summ
mary of Data Colle
ection Re
esults
The overall result of the
t data colleection processs was disapppointing, withh household screener
(telephon
ne) response rates of onlyy about 12 peercent for eacch approach.. In terms off response rattes
alone, thee ABS proced
dures were probably
p
not superior to w
what might bbe achieved in
n a dual fram
me
telephonee survey in Chicago.
C
Lateer we consideer other factoors, includingg the cost of conducting a
dual fram
me telephone survey in a city
c where the telephone nnumbers do not correspo
ond preciselyy to
geograph
hic boundariees, and find some advantaages of the A
ABS sample eeven though rresponse ratees are
comparab
ble.
One of th
he main goals of the pilott was to deterrmine if one of the two aapproaches to
o data collection
was superrior to the otther. The results above aggain find no difference inn response raates for Apprroach
1, which went
w directlyy to the telep
phone if a pho
one number could be maatched to thee address, andd
Approach
h 2, which went
w to mail for
f all househ
holds prior too calling to coonduct the ccontrol card aand
NCVS-1 interview. Th
he original plan of using data collecteed in the Appproach 2 maill instrument to
subsamplle household
ds based on high/low
h
riskk of a victimizzation was nnot implemen
nted due to th
he
lower thaan expected response
r
ratees, but we retturn to exam
mine the utilityy of these daata for other
purposes later in the report.
r
Perhaps the
t most pro
omising resultt from the piilot was the rrelatively goood response rrate to the
Approach
h 2 mail survvey. The maill response ratte was 46 perrcent; this raate is reasonaably good forr a
densely populated,
p
meetropolitan area
a especiallyy since somee response-ennhancing metthods were n
not
used in th
he pilot.
NCVS-CS Pilot
P
Report
2-10
Data Collecttion Results
2
Reviewin
ng the process of data colllection, somee proceduress could be im
mplemented to improve
response rates. For th
he mail surveyy, token inceentives (incluuding $1 or $22 in the initiaal mailing), uusing
special mail
m (FedEx or
o Priority Maail), and adding a third m
mailing of the survey are m
methods that have
a good track record of increasing response
r
ratees. Another iimportant coomponent off the mail surrvey
is the acqquisition of th
he telephone number, and
d some reseaarch into metthods of imp
proving the return
rate for th
he phone number could be
b investigateed. This reseearch would llikely involvee questionnaiire
design (placement and
d text of the telephone reequest) and w
would need ccognitive testting. It shouldd be
noted thaat these meth
hods are morre pertinent to Approach 2 because allll sampled ho
ouseholds aree
mailed in that approacch. It should
d also be noteed that increaasing the maiil response raate would no
ot
necessarilly result in im
mprovementss at the telephone intervieew stage.
Increasing response raates for the telephone
t
surrvey is more problematicc, especially w
without majo
or
changes in
i the telepho
one interview
w content wh
hich was not within the sccope of the p
pilot. With
Approach
h 2, an incentive could bee mailed to th
he householdd prior to thee telephone ccall. A similarr
mailing iss possible in Approach
A
1, but the effect would be diminished ffor addressess that have a
telephonee number thaat is not corrrectly matcheed to the adddress. The size of the inceentive for thiss
componeent is not welll established
d in the literatture, but we suspect a levvel of $5 or $$10 might havve a
better chaance of increeasing the telephone respo
onse rate subbstantially. E
Even then, thee increase in the
telephonee response raate might nott be more thaan 5 to 7 perrcentage poinnts. Post-paidd incentives
generally are not very useful in teleephone surveeys unless thhe amount paaid is consideerably higher and
even then
n the relation
nship is not clearly
c
established.
Another way
w to increaase the teleph
hone response rate mightt be to make changes in tthe content o
of the
interview
w and its nature. One of th
he especially problematic features withh respect to the responsee rate
is the fulll enumeration
n at the begin
nning of the interview. O
Other contentt-related issuues are discusssed
in Section
n 4. A changee that would most substaantially raise tthe responsee rate is a design change th
hat
would allow any adultt to respond to the surveyy for all houssehold membbers. We do not considerr any
of these changes
c
at th
his point becaause they would undoubttedly create m
more differen
nces from thee
core NCV
VS interview
w protocol and
d were essen
ntially out of scope for thee pilot.
NCVS-CS Pilot
P
Report
2-11
Resultss of Oth
her Exp
perimen
nts
3
a Key Fin
ndings
Chapter Highlights and
An experiment ussing different survey titles (one
(
generic, oone using a fo
orm of “Chicag
go”) had little
imp
pact on respon
nse rate to the
e mail survey.
A similar experim
ment in the tellephone surve
ey had no impa
act on respon
nse to the telephone intervie
ew.
A final experimen
nt included the use of biling
gual materialss in the mail ssurvey. The unilingual versio
on
had
d a marginallyy higher response rate than the bilingual, with very few
w returns using
g the Spanish-language questio
onnaire.
While thee primary exp
perimental co
ondition stud
died in the piilot was the uuse of the tw
wo approachees to
data colleection, we alsso included th
hree other exxperiments bbecause standdard data colllection
procedures for ABS surveys are no
ot yet well-esstablished. T
The results off these experiiments are
described
d in this section.
The first secondary-leevel experimeent was an atttempt to dettermine if loccalizing the m
mail instrumeent to
the Chicaago area wouuld improve response
r
ratees. This incluuded using a ccolloquial naame for the M
MSA
(Chicagolland) in the survey
s
title ass well as usin
ng the specifiic regional naame in the letters and
introducttory text. We had also plaanned to tailo
or the surveyy cover using photos of C
Chicago and tthe
surrounding areas. Th
he idea was to
o highlight th
he value of thhe mail surveey to Chicago
o residents an
nd
compare that to a mo
ore generic naational surveyy that had noo specific apppeal to the lo
ocal residentss.
This idea was inspired
d by some off the work in Dillman, Sm
myth, and Chhristian (20099) who suggest the
cover migght play an im
mportant role in highlighting the valuue of the survvey to the loccal population
n.
Some of the original plans
p
for thiss test were reevised prior tto data collecction when co
ognitive testiing
found thaat people exp
pressed a preeference for a more officiial logo and a plain cover on the mail
instrumen
nt. We specuulated that thiis preferencee might have been associaated with thee contents off the
survey beeing about saafety in the neeighborhood
d. Since a majjor part of thhe experimen
ntal condition
n was
conceived
d to use the cover
c
to disccriminate betw
ween the loccal and generric surveys, th
his finding caaused
us to elim
minate some of
o the cover differences and
a use a moore official coover for both
h versions. A
As a
result, wee now call thiis the “Surveey Name” exp
periment, wi th local and generic treattment levels. The
main diffference betweeen the local and generic versions of tthe mail instrruments was the use of th
he
NCVS-CS Pilot
P
Report
3-1
Resu
ults of Other E
Experiments
3
term “Ch
hicagoland” in
n the local in
nstrument. Fo
or example, tthe cover of the local hass the name “T
The
Survey off Chicagoland
d Neighborh
hoods” and th
he generic is just “The Suurvey of Neigghborhoods..”
This expeeriment was tested
t
only for
fo Approach
h 2.
The respo
onse rate forr the local verrsion of the Survey
S
Namee experimentt was 45.3 peercent (s.e.=00.6)
while for the generic version
v
it waas 47.8 percen
nt (s.e.=0.7). The 2.4 percentage poin
nt difference
(s.e.=0.8)) is substantivvely small, buut statisticallyy significant. It also goes in the oppossite direction
n than
anticipateed with the generic
g
versio
on having thee higher resp onse rate.
The secon
nd of these experiments
e
was
w a minor extension off the same iddea to the surrvey introducction,
and is callled the “Survvey Introducction” experim
ment. For thhis experimennt, the teleph
hone interview
wer
used the term
t
Chicago
oland for thee local treatm
ment in the inntroduction aand did not refer to any
geograph
hy for the gen
neric version.. This was tested for bothh approachess, but the treaatments for
Approach
h 2 were defiined by whetther the locall or generic vversion was uused for the m
mail instrument.
For Apprroach 1, the NCVS-1
N
houusehold telep
phone responnse rate was 111.8 percent (s.e.=0.4) fo
or the
local verssion and 12.88 percent (s.e.=0.4) for th
he generic verrsion. In thiss case, the 1 p
percentage point
differencee is not statisstically signifi
ficant (s.e.=0.6). For App roach 2, the NCVS-1 houusehold
telephonee response raate was 12.2 percent
p
(s.e.=
=0.4) for thee local versionn and 12.5 percent (s.e.=0.6)
for the geeneric version
n and the diffference is no
ot significantt.
The final of these exp
periments waas a test of sending responndents eitherr an English--only mail surrvey
or packagge with both English and Spanish survveys. The ideea was to dettermine if sen
nding the surrvey
in both laanguages wouuld improve response rattes for Spanissh-speakers aand whether it would redduce
response rates for mo
onolingual En
nglish speakeers. This expeeriment was conducted o
only for Apprroach
2 and wass restricted to
o householdss that: (1) weere not in lingguistically-isoolated areas ((census tractss that
had moree than 13 perrcent of the population
p
cllassified as sppeaking only Spanish); an
nd, (2) did no
ot
have a suurname that was
w classified
d as Hispanic. All sampledd householdss in the linguuistically-isolaated
areas and
d those with Hispanic
H
surn
names were mailed
m
the biilingual mateerials. A randdom sample o
of
1,500 add
dresses from the Approacch 2 sample in
i other area s was assigneed to get the bilingual
materials and the rem
mainder got th
he English-on
nly survey.
The mail response ratte for the treaatment with the bilinguall materials waas 44.8 perceent (s.e.=1.3) and
for the English-only treatment
t
thee rate was 48.5 percent (s .e.=0.6). Thee 3.7 percenttage point
differencee is statisticallly significant, although itt is not very llarge substanntively. Even
n more surpriising,
only 4 off the 1,500 mail screeners that were mailed the bilinngual materials returned the survey
NCVS-CS Pilot
P
Report
3-2
Resu
ults of Other E
Experiments
3
completeed in Spanish. These resullts are in con
ntrast to findiings from thee National H
Household
Education
n Survey con
nducted by mail
m survey (B
Brick, William
ms and Monttaquila 2011). That surveyy had
no differeence in rates when bilinguual survey material was m
mailed and haad a much higgher rate of
return in Spanish.
Referencces
American
n Association
n for Public Opinion
O
Research (AAPO
OR). 2011. Standard defin
nitions: Finall
disp
positions of case codes and outcome rates for surrveys.
aap
por.org/Conttent/NavigattionMenu/A
AboutAAPOR
R/StandardsampEthics/SStandardDeffinitio
ns/
/StandardDeefinitions2011.pdf
Brick, J. Michael,
M
Willliams, Dougllas, and Mon
ntaquila, Jill. ((2011). Addreess-Based Saampling for
Sub
bpopulation Surveys, Pubblic Opinion Quuarterly, 75, 4409-428.
Dillman, D., Smyth, J., and Christiian, L. (20099). Internet, maail, and mixed--mode surveys: the tailored dessign
metthod, 3rd Edittion. Wiley & Sons, New York.
NCVS-CS Pilot
P
Report
3-3
Co
ost Com
mpariso
ons
4
a Key Fin
ndings
Chapter Highlights and
App
proach 1 was less expensive
e per unit than Approach 2 , whether the unit was com
mpleted households
or completed
c
NC
CVS-2s.
Sub
bsampling bassed on likeliho
ood of victimizzation in Apprroach 2 does n
not appear to reduce per-un
nit
cossts.
Sub
bsampling of unmatched
u
ad
ddresses is lik
kely to reduce per-unit cost; the effect on the variance of
victtimization estiimates was no
ot addressed.
The
e cost of Appro
oach 1 is roug
ghly comparab
ble to a nation
nal RDD usingg both landline
e and cell fram
mes,
dep
pending on the
e treatment of cell-only hou
useholds in thee RDD design.
A successsful NCVS Companion
C
Survey
S
must be
b cost-efficiient. Many asspects of thee pilot study
design sp
pecifically add
dressed cost, including th
he use of maill and telephoone rather than in-person
n
interview
wing and limitting the surveey to a singlee, unboundedd iteration. T
This section w
will compare costs
of mail an
nd telephonee administratiion in the pillot across thee two experim
mental treatm
ments and wiithin
each treattment by oth
her aspects off the design that
t could bee manipulated to improvee efficiency.
In Appro
oach 1, these aspects are whether
w
a telephone num
mber was obtaained from th
he sample veendor
for a sam
mpled addresss, and then whether
w
that number
n
was found not too be associateed with that
sampled address,
a
eitheer because th
he number was
w not in serrvice or was aassociated wiith a differen
nt
address. The
T CS pilot design incluuded sending mail screeneers to these ““bad numberrs,” and then
followingg up by teleph
hone again iff a new numb
ber was obtaained throughh the mail.
Approach
h 2 included other manip
pulatable aspeects besides ttelephone maatch status. SSince all samp
pled
addressess were includ
ded in the maail screener, for
f matched aaddresses theere were maiil responderss with
respondeent-provided numbers, maail responderrs who did n ot provide a telephone number, and m
mail
non-respo
onders. The Approach 2 mail screeneer also includded questionss to allow claassification of
respondeers into “high
h-risk” and “llow-risk” straata based upoon their estim
mated likelih
hood of reporrting
a victimizzation in the telephone fo
ollow-up inteerview.
NCVS-CS Pilot
P
Report
4-1
Cost Co
omparisons
4
Generallyy, survey costts for a given
n design mayy be classifiedd as either fixxed or variablle. Variable ccosts
are those directly prop
portional to the
t sample siize. For the ppurposes of tthis report, w
we will focus only
on compaaring variablee costs acrosss treatments and other deesign aspectss within the C
CS pilot. Our
general ap
pproach is to
o derive component unit costs based oon observed data, and then to incorpo
orate
these into
o models that would supp
port explorin
ng what woulld happen wiith certain ch
hanges in the
design. The
T output off the models is cost per so
ome output uunit, such as completed ttelephone
interview
ws where the respondent
r
reports
r
at leaast one victim
mization.
4.1
Mail Costs
C
Table 4-11 presents thee cost modell results for th
he Approachh 1 mail screeener. Table 44-1 includes tthree
separate groups:
g
addresses withouut a matched telephone nuumber, all off which were sent mail
screeners; addresses with
w a matcheed telephonee number, noone of which were initiallyy sent a mail
screener; and matched
d addresses where
w
the telephone num
mber proved ““bad,” which
h were
subsequently sent a mail
m screener. Note that th
hese last are a subset of alll matched saample addressses.
Table 4-1
1.
Approach 1 mail screener
s
costts
Match
Initial sample
Match ra
ate
Subsamp
pling rate
Initial ma
ailing
Comp
pletion rate
Comp
pleted surveyss
Telephon
ne rate
Returrns after 2 weeks
Relative cost per mail
m complete
Relative cost per te
elephone case
e
Relative cost per person
w/vicctimization
Including
g cases not ma
ailed to
12,500
0
41.1%
%
74.1%
%
40.0%
%
Bad
d#
A
All
65
58
0.4
43
28
80
20.4
4%
5
57
4
45
8.1
1%
1.5
57
5.5
54
5,14
42
3.7
73
No m
match
7,3
358
1
1.00
7,3
358
33
3.0%
2,4
428
1,8
800
13
3.2%
0
0.99
3
3.71
0
0.97
All
7,6
638
2,4
483
1,8
845
1
1.00
1
1.00
1
1.00
0
0.58
There aree three comp
ponents of vaariable cost fo
or the mail sccreener: (1) tthe initial maailing and
reminder postcard; (2) the second mailing; and
d (3) receipt aand processinng of returneed questionn
naires.
To calcullate the total cost of the mail
m screener effort for a column, we multiply per--unit costs fo
or
each of th
hese compon
nents times th
he number of
o cases to whhich it is appplicable, and ssum the resuults.
Initial maailing cost is multiplied
m
byy the initial saample; seconnd mailing coost is multipliied by the
number of
o cases not returned
r
afteer 2 weeks; an
nd receipt annd processingg cost is multtiplied by thee
NCVS-CS Pilot
P
Report
4-2
Cost Co
omparisons
4
number of
o returns. To
o obtain totaal per-unit co
osts for outcoomes, we divvide the total cost for the
column by
b the numbeer of cases with
w the outco
ome of intereest.
meters in thee model, whiich may be ad
djusted, incluude the match rate, the suubsampling rrate,
Key param
the comp
pletion rate, and
a the teleph
hone rate (th
he proportionn of returnedd screeners w
with telephon
ne
numbers)). Note that some
s
derived
d numbers in
n the table maay not equatee exactly to tthe actual pilo
ot
results beecause of rouunding.
The last four
f
rows of Tables 4-1 summarize th
he outcome uunit costs forr the Approaach 1 screeneer,
using the cost for all Approach
A
1 cases
c
as a ben
nchmark. Thhe cost per coompleted maail screener w
was
about 60 percent high
her for the ad
ddresses with
h “bad” matcched numberrs than for th
he non-match
hed
addressess. The relativve costs per case sent for telephone
t
intterviewing w
were the samee as for returrned
screeners because onlly cases with telephone nuumbers weree sent on, andd the rate of including
numbers was assumed
d to be the saame across columns. Thee cost per com
mpleted indiividual teleph
hone
interview
w with a reporrted victimization (denom
minator exterrnal to the tabble) was morre than 3.5 tiimes
higher for the “bad” numbers,
n
butt this estimatte is unstablee because of tthe small sam
mple size. Th
he last
row facto
ors in the matched cases that
t were nott in the “badd number” grroup. The rellative cost is
almost th
he same as 1 minus
m
the match rate; thee difference iis because off the “bad nuumber” casess.
Table 4-22 presents esssentially the same
s
inform
mation for Appproach 2. Heere, the only split is betw
ween
matched and unmatch
hed addresses, and all maatched addressses were avaailable for tellephone follo
owup, regard
dless of the mail
m screenerr outcome. The
T benchmaark for calculaating relativee cost is again
n the
per-unit cost
c of all thee Approach 1 mail cases. The compleetion rate for non-matcheed cases was
virtually identical
i
betw
ween approacches (Table 4-1
4 versus Taable 4-2), as w
were the relaative costs peer
mail returrn and per caase available for telephon
ne follow-up. The cost peer telephone iinterview
reportingg victimization was higherr for reasons we will see llater.
For Apprroach 2 matcched cases, th
he completion rate was hiigher than foor unmatchedd cases, whicch is
not surprrising, so the relative cost per mail com
mplete is aboout 70 percennt of that forr the unmatch
hed
cases. An
nd, since all matched
m
cases except thosse identified as nonresideential or vacaant from posttal
returns were
w availablee for telephon
ne follow-up
p, the cost perr case availabble for the teelephone wass only
about 28 percent of th
hat for unmaatched cases. But, of courrse, the mail cost for all m
matched cases
with App
proach 1 is much
m
lower sttill, because only
o those wiith bad numbbers were maailed to.
Continuin
ng to the nexxt-to-last row
w of Table 4-2, we see thaat the relativee mail cost peer person
reportingg victimization in the telep
phone intervview was mucch higher thaan per teleph
hone case, agaain
for reason
ns we will leaarn later.
NCVS-CS Pilot
P
Report
4-3
4
Cost Co
omparisons
Table 4-2
2.
Approach 2 mail screener
s
costts
Match
Initial sample
Match ra
ate
Subsamp
pling rate
Initial ma
ailing
Comp
pletion rate
Comp
pleted surveyss
Telephon
ne rate
Returrns after 2 weeks
Relative cost per mail
m complete
Relative cost per te
elephone case
e
Relative cost per person w/ victimization
Including
g Approach 1 cases
c
not maiiled to
14,000
0.398
0.75
40.0%
5,572
1.0
5,572
47.0%
2,619
1,964
18.8%
0.71
0.93
0.76
No mattch
8,42
28
1
1.0
8,42
28
33.0
0%
2,78
81
2,08
86
13.2
2%
0.9
99
3.6
67
0.9
97
A
All
14,0
000
14,0
000
38
8.6%
5,4
400
4,0
050
0
0.85
1
1.67
0
0.88
1
1.52
Overall, the
t total maill cost per perrson reportin
ng victimizatiion in the tellephone interrview was jusst
over 50 percent
p
higheer in Approacch 2 than in Approach
A
1,, when factorring in the m
matched casess that
were not mailed to in Approach 1. This differeence was exppected, since the sample ssize was largeer in
Approach
h 2, anticipatting subsamp
pling for telep
phone follow
w-up, and sinnce all Approach 2 sampleed
addressess were includ
ded in the maail portion off the pilot tesst. Our hypotthesis was th
hat this cost
differencee would be compensated
c
for by savin
ngs in the teleephone follow
w-up, and/o
or by the valuue of
the additiional data ob
btained in Ap
pproach 2.
4.2
Teleph
hone Cossts
The cost model for th
he telephone portion of th
he CS is buillt somewhat differently frrom the maill
or each colum
mn of the co
ost tables (Taables 4-3 and 4-4) and sepparately for th
he householdd
model. Fo
screener and
a for the victimization
v
screener and
d associated iincident repoorts, we deveeloped an
algorithm
m that incorpo
orated the mean
m
length of
o the intervieew and the nnumber of caalls across all cases
in the cattegory divided by the num
mber of comp
pleted intervviews to estim
mate a total n
number of
interview
wer minutes per
p completed
d interview. This
T result w
was multipliedd by a mean cost per
interview
wer hour for all
a marginal activities
a
asso
ociated with ttelephone daata collection
n and data
preparatio
on. The algorithms were adjusted so that
t the totall cost across all cases mattched the acttual
incurred marginal
m
cossts.
t virtually all Approach
h 1 cases withh a telephonee number weere sent for
In Table 4-3, we see that
we received a
telephonee follow-up. The only “suubsampling” was from noon-matched ccases where w
mail screeener with a teelephone num
mber too latee to follow uup. The household screen
ner completio
on
NCVS-CS Pilot
P
Report
4-4
Cost Co
omparisons
Table 4-3
3.
4
Approach 1 teleph
hone and tota
al costs
Ma
atch
Available
e for telephone
e
Subsamp
pling rate
Sent for telephone
t
folllow-up
Househo
old screeners
Comp
pletion rate
Numb
ber completed
d
Victimiza
ation screenerrs
Comp
pletion rate
Numb
ber completed
d
Relative telephone
e cost per
Relative total cost per
Persons reporting victiimization
Reporting rate
Numb
ber reporting
Relative telephone
e cost per
Relative total cost per
Property crimes
Rate pe
er HH screene
er
Numb
ber completed
d
Relative telephone
e cost per
Relative total cost per
Personall crimes
Rate pe
er HH screene
er
Numb
ber reporting
Relative telephone
e cost per
Relative total cost per
Bad #
45
1.00
45
All
5,142
0.99
5,097
No matcch
1,800
0
0.98
8
1,767
7
A
All
6,9
987
0
0.99
6,9
909
23%
11
13%
657
34%
%
593
3
1
18%
1,2
261
1.18
13
0.59
1.98
1.19
782
1.25
0.80
1.16
6
687
7
0.72
2
1.21
1
1
1.18
1,4
482
1
1.00
1
1.00
15%
2
1.00
3.39
21%
142
1.54
0.98
32%
%
191
1
0.59
9
0.99
9
2
26%
3
335
1
1.00
1
1.00
0%
0
N/A
N/A
26%
169
1.49
0.95
36%
212
0.60
1.01
30
0%
38
81
1.0
00
1.0
00
9%
1
0.40
1.36
5%
33
1.64
1.05
8%
48
0.57
0.96
7%
8
82
1.0
00
1.0
00
rate was higher
h
for those cases thaat came from
m the mail scrreener than ffor those for which we sim
mply
called thee matched nuumber, but of course therre was substaantial mail noonresponse fo
for the match
hed
cases (Table 4-1).
Table 4-33 presents completion or reporting rattes, and relatiive unit costs, for four diifferent stagees of
the survey process: co
ompleting thee victimizatio
on screener, completing a victimizatio
on screener in
n
which at least one incident was rep
ported (regarrdless of how
w it was codeed later), repo
orted inciden
nts
that weree coded as property crimees by NCVS rules,
r
and repported incideents coded as personal crrimes.
Note thatt there are 3 different levels representted in Table 44-3, which arre also shown
n in Table 4--4 for
Approach
h 2: address or
o household
d, person, and incident. T
Tables 4-3 annd 4-4 differ iin this way frrom
the tabless in Section 2,
2 which stay at the address or househhold level.
NCVS-CS Pilot
P
Report
4-5
Cost Co
omparisons
4
The comp
pletion rate for
f victimizattion screenerrs (NCVS-1) is greater thhan 1 in mostt cases becauuse
more than
n 1 adult couuld be sampleed in a houseehold. As nooted in Sectioon 2, the victiimization
screener completion
c
rate
r for the household
h
scrreener responndent was m
much higher tthan that for
other aduults sampled in
i a househo
old. The relattive telephonne cost for vicctimization sscreeners is
highest fo
or the matched cases, as it
i is directly related
r
to thee completionn rate. But, faactoring in m
mail
costs, thee cost per vicctimization sccreener in thee matched saample was abbout 2/3 thatt of the
unmatcheed sample. (T
The total cost for the mattched Appro ach 1 cases iincludes mailling a
“prenotiffication” letteer.) In each cost
c row, the benchmark cost is the coost across all of the Apprroach
1 sample..
Individuaals in the non
n-matched saample were about 50 perccent more likkely to reportt a victimizatiion
than thosse in the matcched sample,, so that the “per-victim”” cost is virtuually identicall between
matched and non-mattched cases. Matched cases whose venndor-provideed telephonee number was not
in servicee or not assocciated with th
he sampled address
a
and ffor which a m
mail screenerr was returned
with a diffferent teleph
hone numberr were by farr the most exxpensive in A
Approach 1, b
but the samp
ple
size is sm
mall, so the co
ost estimates are relativelyy unstable.
The cost ratios for rep
ported propeerty and personal crimes aare very simiilar to those ffor victimizaation
screeners with any rep
ported incideent. Again, beecause of thee small sample size, the raatios for the “bad
number” cases are un
nstable.
Table 4-44 presents thee same inform
mation for Approach
A
2, bbroken out bby more samp
ple groups. SSince
all sampleed addresses were sent mail
m screeners, all matchedd addresses w
were availablee for telephone
follow-up
p but could have
h
(1) a maail screener with
w a telephoone number, (2) a mail screener witho
out a
telephonee number, an
nd (3) no maiil screener. All
A Approachh 2 cases withh a completedd mail screen
ner
were assiggned to a riskk stratum based on their responses. B
By design, aboout 25 percent were assiggned
to the “hiigh risk” straatum, indicatiing a higher predicted
p
likkelihood of reeporting a vicctimization in
n the
telephonee interview. The
T proportiion of high riisk cases variied by samplle subgroup, with the
unmatcheed addresses having the laargest propo
ortion (30 perrcent). Exceppt for the maatched mail n
nonrespondeers, all availab
ble cases weree sent for tellephone folloow-up. “Subssampling” included 2-4
percent of
o mail screen
ners received
d too late for follow-up, aand 50 percennt subsampling of mail
nonrespo
onders.
NCVS-CS Pilot
P
Report
4-6
Approach
A
2 telep
phone and total costs
4
Mail
M #
4-7
R
Risk proportion
A
Available for telep
phone
S
Subsampling rate
S
Sent for telephone
e follow-up
H
Household screeners
Completion ratte
Number completed
V
Victimization screeners
Completion ratte
Number completed
Relative teleph
hone cost per
Relative total cost
c
per
P
Persons reporting victimization
Reporting rate
Number reportting
Relative teleph
hone cost per
Relative total cost
c
per
P
Property crimes
Rate per HH scre
eener
Number completed
Relative teleph
hone cost per
Relative total cost
c
per
P
Personal crimes
Rate per HH scre
eener
Number reportting
Relative teleph
hone cost per
Relative total cost
c
per
Match
Mail no #
High
Low
0.223
0
0.777
148
516
0.95
0.97
141
500
No m
mail
No
o Match
M
Mail #
High
Low
0.30
0.70
638
1,480
0.96
0.97
612
1,430
7,690
0.78
6,012
High
0.228
460
0.97
445
Low
0.772
1,555
0.97
1,507
30%
134
27%
407
9%
13
10%
52
6%
%
78
8
33%
200
33%
466
23%
1,350
1.12
150
0.91
0.76
1.21
493
0.72
0.66
1.00
13
1.96
1.99
1.06
55
1.60
1.63
1.03
3
80
0
2.86
6
4.40
0
1.18
236
0.82
1.29
1.18
550
0.67
1.20
1.17
1,577
0.89
1.18
61%
91
0.40
0.33
13%
62
1.52
1.38
69%
9
0.75
0.76
15%
8
2.91
2.96
23%
%
18
3.36
6
5.15
15
56%
133
0.39
0.60
16%
86
1.13
2.02
407
0.91
1.21
82%
110
0.32
0.27
13%
53
1.73
1.57
38%
5
1.31
1.33
6%
3
7.55
7.70
32%
%
25
5
2.3
36
3.6
61
77%
154
0.32
0.50
14%
65
1.46
2.57
31%
415
0.87
1.15
11%
15
0.50
0.42
4%
15
1.32
1.19
8%
1
1.41
1.43
4%
2
2.44
2.48
4%
%
3
4.2
23
6.4
48
17%
33
0.33
0.50
2%
9
2.27
4.00
6%
78
0.99
1.32
N/A
A
2,893
3
0.48
8
1,37 7
Total
Cost Comparisons
C
C
i
NCVS-CS Pilot Report
TTable 4-4.
Cost Co
omparisons
4
Overall, the
t householld screener co
ompletion raates were com
mparable to tthose of App
proach 1, with
h the
lowest rattes for match
hed cases thaat either did not
n respond to the mail sscreener or reesponded buut did
not proviide a telephone number. The
T complettion rate for vvictimizationn screeners w
was also
comparab
ble to that off Approach 1,
1 with the saame sample ssubgroups haaving the low
west rates as ffor
the houseehold screeneer.
The overall Approach
h 1 costs are used as bencchmarks for eeach cost row
w. As with A
Approach 1, tthe
relative teelephone cosst per victimization screen
ner is inverseely proportioonal to the co
ompletion ratte.
Overall, telephone
t
costs were som
mewhat lowerr for Approaach 2, but this difference iis entirely duue to
subsamplling mail non
nresponders. The total co
ost per victim
mization screeener, includin
ng mail cost, is
higher ovverall than for Approach 1 for each off the unit cosst parameterss – about 20 percent high
her
for victim
mization screeeners, victim
mization screeeners with anny reported inncident, and reported
property crimes, and more
m
than 300 percent higgher for repoorted personaal crimes. Th
hus, the prediicted
benefits of
o conductingg the mail screener with all
a sampled aaddresses didd not materialize. The maiil
nonrespo
onders were by
b far the mo
ost expensivee in Approacch 2 for 3 of the 4 unit co
osts, and were
between 3 and 6 times as expensivve as the App
proach 1 bennchmark.
As discusssed in Sectio
on 8, the maiil screener did
d reasonablyy well at preddicting wheth
her any incideent
would bee reported, buut 13 to 16 percent of resspondents in low risk houuseholds also
o reported at least
one incid
dent. The ratee of reported
d property criimes was aboout 6 times hhigher in high
h risk househ
holds
than in lo
ow risk houseeholds overalll, and the rate of reporteed personal ccrimes was ab
bout 5 times
higher.
mization varieed widely am
mong the Appproach 2 sam
mple groups. F
From
Cost per person reporrting a victim
t
the cosst of the App
proach 2 ben
nchmark. Sim
milar variationn may be obsserved in the
1/3 to 5 times
relative co
osts per prop
perty crime and
a per perso
onal crime reeported. This variation waas a function
n of
the differrences in repo
orting rates, as well as in other factorss discussed eearlier.
4.3
Relatio
onship off Mail an
nd Teleph
hone Cossts
As shown
n in the previious section, the total cosst for Approaach 2 was abbout 20 perceent higher than
that for Approach
A
1 for
f three of th
he unit costss (per victimiization screenner, per perso
on with a
reported incident, and
d per propertty crime) calcculated, and aabout 30 perrcent higher p
per reported
personal crime. Howeever, the telep
phone data collection
c
cosst in Approaach 2 was abo
out 10 percen
nt
NCVS-CS Pilot
P
Report
4-8
Cost Co
omparisons
4
lower thaan Approach 1 for the firsst three unitss, and about the same perr reported peersonal crimee.
The reaso
on for the low
wer telephon
ne costs was the subsamppling of nonrespondents tto the mail
screener. Had that sub
bsampling no
ot occurred, the telephonne costs woulld have been
n about the saame
t two apprroaches for th
he first threee units, and aabout 12 perccent higher fo
for Approach
h2
between the
per reporrted personall crime. With
hout the subssampling, Appproach 2 woould have beeen slightly more
expensivee per unit thaan with it.
The reaso
on for the ovverall differen
nce in cost beetween the tw
wo approachhes was in thee cost of the mail
screener, which, as the design wass implementeed, representeed 57 percennt of the totall cost for
Approach
h 2, as compared with 38 percent of the
t total costt for Approacch 1. Had thee subsamplin
ng
not occurrred, mail costs would stilll have accouunted for aboout 53 percennt of total co
osts for Apprroach
2.
4.4
Effect of Chang
ging Para
ameters
The mod
del allows us to
t estimate th
he effect on cost of channges in param
meters, such aas subsamplin
ng
rates, resp
ponse rates, and
a some co
ost parameterrs. This sectioon will discuss several off these possib
ble
changes.
One way to increase the
t mail response rate wo
ould be to incclude a smalll cash or otheer incentive iin the
first mailiing. Of coursse, this also increases the cost. A $1 inncentive withh all initial mailings wouldd
have to effect about a 5 percentagge point increease in the m
mail return acrross the boarrd rate to payy for
A
2, and
a 8 points iin Approachh 1. Similarly,, the break-evven
itself in teerms of mail returns in Approach
point for a $2 incentivve would be about a 9 po
oint increase iin response iin Approach
h 2. How this
change would
w
affect the
t telephonee follow-up is not clear. T
The most exppensive casess in Approach 2
are mail nonresponde
n
ers. If househ
holds respond
ding to the inncentive conndition but no
ot otherwise
continue their cooperrative behavio
or into the teelephone folllow-up, a muuch lower inccrease in maiil
response would pay for
fo itself.
Given thee relatively grreater per-un
nit cost of some sample ggroups for thhese survey deesigns, it mayy be
more effiicient to subssample, eitheer before the initial mailinng or based oon the mail sccreener resullts
before tellephone follo
ow-up. For example,
e
the original desiggn called forr subsamplingg low-risk
respondeers in Approaach 2, but thaat plan was abandoned giiven the loweer-than-expected mail
screener response.
r
Insstead, mail sccreener nonrresponders w
with vendor-pprovided teleephone numb
bers
NCVS-CS Pilot
P
Report
4-9
Cost Co
omparisons
4
were subssampled in Approach
A
2. Here
H are the cost implicaations from thhe model of subsamplingg at
different points:
Subsampling non-mattched cases before
b
mailinng in Approaach 2 would h
have a very ssmall
odel shows a slight decreaase in the cosst per victimiization screeener
effect on cost; the mo
and a som
mewhat largeer, but still sm
mall increase in the cost pper person wiith reported
victimizattion, becausee the non-maatched cases are more likeely to report a victimizatiion
than the matched
m
casees.
Subsampling non-mattched cases before
b
mailinng in Approaach 1 would n
not be effecttive.
Howeverr, mailing to the
t matched “bad numbeer” cases doees not seem tto be costeffective, although as we have notted the samp le size is relaatively small.
ders in Approoach 2 appeaars to have haad a slightly
Subsampling the maill nonrespond
positive effect
e
on variious unit cossts.
Subsampling mail resp
ponders classified as “low
w risk” wouldd be close to
o cost neutrall (the
model sh
hows very sligghtly positivee, but undoubbtedly well w
within the maargin of errorr) for
matched cases where a telephone number was not provideed in the retuurned mail
screener, but not for the
t other two
o low risk suubgroups (maatched with m
mail phone
number and
a nonmatcched).
A significcant caveat with
w these obsservations ab
bout efficienccy is that theey do not con
nsider the efffect
of subsam
mpling on thee variance off survey estim
mates. Since aany differenttial sampling would likelyy
increase the
t variance and
a hence reeduce the effe
fective samplee size, of thee possibilitiess explored ab
bove
only drop
pping the maail screener fo
or matched “bad
“
numberr” cases for A
Approach seeems promisin
ng
within the parameterss of the desiggns as implem
mented in thee pilot.
Another significant caaveat is that all
a of these observations
o
aare specific tto the design
ns implementted,
and to their performaance in one siite. The pred
dictive powerr of the cost m
models is thuus fairly limitted;
changes in
i the design or implemen
ntation in oth
her sites migght produce vvery differentt results.
4.5
Compa
arison wiith RDD Costs
C
Since a raandom-digit-dial (RDD) survey
s
would
d be another possible appproach to an NCVS
Companiion Survey deesign, we willl attempt to compare thee cost of the ppilot to that of an RDD
survey wiith a similar design
d
and siimilar conten
nt. Since the ppilot did nott include an R
RDD samplee, we
will draw upon Westaat’s experiencce with an RD
DD survey thhat was in thhe field at thee same time aas the
CS pilot, in Californiaa.
NCVS-CS Pilot
P
Report
4-10
Cost Co
omparisons
4
One of th
he major issuues in consideering an RDD
D design is ccoverage of hhouseholds w
with only celllular
telephonee service. The most recen
nt estimates from
f
the Nattional Healthh Interview Survey indicatte
that 34 peercent of aduults nationallyy live in cell-only househ olds; the prooportion is m
more than 50
percent for
fo adults und
der 35 (Blum
mberg and Luuke, 2012). M
Most new RDD surveys no
ow include a
sample off numbers drrawn from th
hose assigned
d to cellular sservice. Thesse surveys maay or may no
ot
screen ho
ouseholds reaached througgh those num
mbers to interrview only inn cell-only ho
ouseholds. Frrom a
data colleection perspeective, it is more efficient to interview
w in all eligiblee householdss reached thrrough
the cell saample.
A furtherr complicatio
on with an RD
DD design fo
or a local areea survey is thhat cell numb
bers are assiggned
based on where a telephone is purrchased ratheer than wheree the purchaser lives, andd that once
assigned they
t
tend to be more porrtable than laandline numbbers. Thus, inn sampling fo
or a local area,
identifyin
ng the cellularr area code-eexchange com
mbinations too include is leess precise th
han for a landline
sample. Certainly
C
for a cell samplee, and depend
ding on the aarea for a lanndline samplee as well, the data
collection
n process muust include so
ome screenin
ng to ensure tthat only resiidents of the target area aare
included in the surveyy estimates. Such
S
screenin
ng requires thhat the surveey ask about tthe respondeent’s
residentiaal address; ineligible households may be
b screened oout before or after conduucting the
interview
w. Thus, a resiidential eligib
bility rate muust be factoreed into the coost model for a local areaa
RDD surrvey.
The 20111 California Health
H
Intervview Survey (CHIS)
(
was iin the field att the same tim
me as the CSS
pilot. CH
HIS is a dual-fframe (landlin
ne and cell) RDD
R
surveyy of Californiia households with a desiggn
somewhaat similar to that
t of the piilot: after a veery brief screeening intervview with a ho
ousehold
respondeent, one random adult is selected
s
for a longer interrview; in houuseholds with
h children (0--11)
and/or ad
dolescents (112-17), one ch
hild and/or one
o adolesceent is also sellected. The content of thee
interview
ws is health co
onditions, heealth care, heaalth insurancce, and relateed topics. CH
HIS 2011 resp
ponse
rates are not
n yet availaable, but gen
nerally the scrreener responnse rate is higgher than thaat of the CS pilot,
and the conditional reesponse rate for the longeer interview iis somewhat lower.
Obviouslly there are many
m
differen
nces to take into account when compaaring two surrveys done w
with
different sampling fraames on diffeerent topics in different ggeographic arreas. Thus, an
ny cost
comparison can only be very broaad. With these caveats, ouur comparisoon of CHIS 22011 costs wiith
the CS piilot indicate that
t the telep
phone-related
d cost per CSS household w
with at least one victimizzation
screener is
i about 58 percent
p
of thaat for one co
ompleted CH
HIS adult inteerview. Given
n that in Secttion
4.3 we fo
ound that teleephone intervviewing costss accounted ffor about 62 percent of tthe total costt per
NCVS-CS Pilot
P
Report
4-11
Cost Co
omparisons
4
victimizattion screenerr for Approach 1, we can say that CS Approach 1 and CHIS costs are closee to
equal for close to the same kind of result.
hone numberrs dialed weree geographicaally eligible, ssince we werre covering th
he
In CHIS,, most teleph
entire state. In a smalller area or on
ne proximatee to other largge population centers (wh
hich Californ
nia is
not), we would
w
likely encounter hiigher rates off geographic ineligibility, and hence h
higher costs p
per
completeed interview in
i an eligible household. The
T CHIS saample designn called for ab
bout 20 percent
of all inteerviews to bee conducted with
w the cell phone samp le. An optim
mal design forr most areas in
the Uniteed States wouuld probably have a higheer proportionn of sample ffrom the cell frame, which
again wouuld increase the
t cost per completed in
nterview in ann eligible houusehold.
Referencce
Blumbergg, S.J., and Luuke, J.V. Wirreless Substittution: Early Release of E
Estimates fro
om the Nation
nal
Heealth Interview
w Survey, Jan
nuary–June 2012.
2
http
p://www.cd
dc.gov/nchs/
/data/nhis/earlyrelease/w
wireless2012112.PDF.
NCVS-CS Pilot
P
Report
4-12
Quesstionna
aire Asssessme
ent
5
a Key Fin
ndings
Chapter Highlights and
In the
t mail scree
eners, item-levvel nonresponsse was low exxcept for telep
phone numberr (with about 2
25
perrcent missing data).
In the
t Approach 2 mail screener, there is evvidence that soome respondeents had difficculties with the
hou
usehold enum
meration matrixx.
The
ere were some
e differences in
i household composition
c
a
as reported in the Approach
h 2 mail screener
and
d telephone in
nterview.
In the
t telephone enumeration,, household member
m
date oof birth had reelatively high n
nonresponse.
Havving the house
ehold enumerration and dem
mographics att the start of the interview increased the need
for refusal avoida
ance tactics by
b interviewerss.
ns.
Inte
erviewers repo
orted little diffficulty with the
e NCVS1 screeening question
“Re
e-starting” inte
erviews was difficult in a tellephone enviroonment wheree cases are sh
hared (this occcurs
whe
en an interview is complete
ed in more tha
an one session
n).
Inte
erviewers and respondents had some diffficulties unde rstanding the relationship b
between the
NCV
VS1 and NCVS
S2 questions. A particular problem
p
was w
with NCVS2 qu
uestion 88 (theft).
Inte
erviewer had difficulty
d
with questions drivving major skiip patterns in the NCVS2, in
ncluding questtion
10 (location) and
d question 20a
a (presence).
Beh
havior coding indicates thatt the interview
wers usually reead the questiions as worded.
5.1
Overview and CATI
C
Instrument TTiming
The NCV
VS-CS Pilot used
u
two versions of a maail screener aand three CA
ATI questionnaires. The m
mail
screeners are included
d in Appendiix A.
The sole purpose of the
t Approach
h 1 mail screeener was to oobtain a telep
phone numb
ber
for follow
w-up. It incluuded some op
pinion questiions on neighhborhood saafety and poliice
performaance, and onee question on
n tenure (how
w long resideed at current address).
NCVS-CS Pilot
P
Report
5-1
Qu
uestionnaire A
Assessment
5
The Approach 2 mail screener inccluded all of tthe Approach 1 question
ns plus questiions
on victim
mization and expected
e
corrrelates of vicctimization.
The CAT
TI household
d screener inccluded standaard questionss to verify th
hat we had
reached a residence att the sampled
d address andd identify an appropriate household
respondeent, and a set of householld enumeratioon and demoographic queestions from the
NCVS Co
ontrol Card.
The victim
mization screeener includeed questions from the NC
CVS-1, omittting those asking
about ideentity theft.
The incid
dent report in
ncluded all off the questionns in the NC
CVS-2.
The CAT
TI instrumentts were adapted from thee paper versioons formerlyy used by the Census Bureeau.
Westat diid not have access
a
to speccifications fo
or the CAPI pprogram currrently in use on the core
NCVS.
The houssehold screen
ner was comp
pleted with 2,611
2
eligible households,, with a mean
n administrattion
time of 4.8 minutes. There
T
was a fairly
f
wide raange of adminnistration tim
mes, between
n 2 and 25
minutes, depending on
o the size off the househo
olds and wheether the inteerviewer had to continue
persuadin
ng the respon
ndent to partticipate.
A total off 2,991 respo
ondents comp
pleted the vicctimization sscreener; the mean admin
nistration tim
me
was 8.1 minutes,
m
with
h about 90 peercent of resp
pondents com
mpleting it inn between 5 aand 13 minutes.
For thosee reporting a victimization
n, the mean time
t
was 10. 6 minutes, annd 7.3 minuttes for those not
reportingg a victimizatiion.
A total off 1,474 incideents were rep
ported by 890 respondennts (excludingg completed iinterviews w
with
individuaals outside thee Chicago SM
MSA). Mean administratioon time for aall incidents eexcluding
duplicates, those outside the refereence period, and incompllete reports w
was 10.9 min
nutes. Ninetyy
percent of
o all incidentt reports werre completed in between 44.5 and 21 m
minutes. For iincidents codded
as person
nal crimes (co
odes 1-23), th
he mean was 16.2 minutees, and for prroperty crimees (codes 24--41)
10.1 minuutes. Just oveer one-third of
o responden
nts completinng an incidennt report com
mpleted moree
than one.. For the firstt (or only) in
ncident report, the mean aadministratioon time was 111.7 minutes, and
for the seecond inciden
nt report 9.6 minutes. To
otal time spennt on all inciddent reports by one
respondeent ranged fro
om 2 minutes to more than 100, with a mean of 14.7 minutes.
NCVS-CS Pilot
P
Report
5-2
Qu
uestionnaire A
Assessment
5.2
5
Item Nonrespo
N
onse and Responsse Frequencies
Frequenccies for closed
d-ended questions in botth the mail annd telephonee questionnaiires will be
provided separately to
o BJS. This section will diiscuss highligghts of item nnonresponsee and a few
unweightted frequenciies of interestt.
Mail Screener
The two mail screenerr versions haave questionss 1-8 in comm
mon, labeledd “Your Neigghborhood” iin
Approach
h 2. The rem
maining 4 item
ms in Approaach 1 also apppear in Apprroach 2, but h
have interven
ning
questionss. Approach 2 includes quuestions on victimization
v
and correlattes of victimiization, incluuding
householld size, emplo
oyment, whether own or rent, and nuumber of movves in the paast 5 years. Fo
or
the mail screener
s
item
m nonresponsse means thaat there was nno response m
marked or th
hat it was nott
possible to
t determine which respo
onse was marrked.
Nonresp
ponse to Qu
uestions Com
mmon to Ap
pproach 1 an
nd Approach
h 2. Item no
onresponse to
o all
of the Ap
pproach 1 quuestions was less
l than onee percent, exccept for telepphone numb
ber (26% missing).
The telep
phone numbeer was missin
ng in 25 percent of the Appproach 2 sccreeners. In tthe Approach
h2
instrumen
nt, nonresponse to Questtions 1-6 (all on the first ppage and ideentical to the Approach 1
instrumen
nt in wordingg and layout)) was somew
what higher buut still no moore than 1.3 percent.
Question
ns 7 and 8 had
d virtually identical nonreesponse betw
ween the twoo instrumentss, and
nonrespo
onse to Questions 17 and 18 in Appro
oach 2 (policee performancce and 911 reesponse,
Question
ns 9 and 10 in
n Approach 1)
1 was less th
han 1 percentt. However, nnonresponsee to Question
n 23
in Approach 2 (“How
w long have you
y lived at th
his address?””, Question 111 in Approaach 1) was 4.44
percent.
Compariison of responses in Ap
pproach 1 an
nd Approach
h 2. The disttribution of rresponses to
common questions was
w very similar between Approach
A
1 aand Approacch 2. For the first 8 questiions
in each in
nstrument, Approach
A
2 reespondents gave
g
slightly m
more positivve responses tthan Approaach 1
respondeents. Question
n 17 in Apprroach 2 (police performannce) would bbe most likelyy to show a
context effect,
e
as it im
mmediately fo
ollowed the victimization
v
n questions. A
Approach 2 rrespondents w
were
about 3 percentage
p
po
oints more likkely than Ap
pproach 1 resspondents too “strongly aggree” that thee
police aree doing a goo
od job, and about 4 pointts less likely tto disagree orr strongly dissagree. This
differencee was the larggest observed among thee common iteems, and sugggests that there might bee a
NCVS-CS Pilot
P
Report
5-3
Qu
uestionnaire A
Assessment
5
very smalll context efffect arising frrom the inserrtion of the vvictimizationn items. Not ssurprisingly, those
who repo
orted a victim
mization weree far more likkely (21% to 4%) to disaggree or stronggly disagree tthat
the policee are doing a good job.
Nonresp
ponse to queestions onlyy in Approacch 2. The vicctimization qquestions (9-116) all had iteem
nonrespo
onse of less th
han 1 percen
nt. Question 19 is a 5x2 gr
grid asking foor the numbeer of males an
nd
females in
n the househ
hold by age raange; 2.5 perrcent of respoondents did nnot enter a n
number greatter
than 0 in any of the boxes. More than
t
2 percen
nt of responddents did nott answer Queestion 20, ab
bout
unemployyment. Quesstion 21 requiires a “yes” or
o “no” to fivve employmeent fields; 7 p
percent of
respondeents left the first
fi item blan
nk, and each of the other items had 122 to 14 perceent nonrespo
onse.
On the laast page of th
he Approach 2 booklet, Question
Q
22 hhad just overr 2 percent no
onresponse.
Question
n 24 (numberr of moves in
n the past 5 years)
y
was thee only item inn the survey that could haave
been skip
pped approprriately, if the response to Question 233 was 5 years or longer. M
More than 3
percent of
o respondents did not an
nswer either Question
Q
23 or Question 24, but almo
ost none skip
pped
Question
n 24 after giviing an answeer of less than
n 5 years to Q
Question 23. Of those wh
ho said “5 orr
more years” in Questiion 23, 34 peercent went on
o to answerr Question 244. Of those, m
most (95%) ggave
the expeccted answer of
o “0.”
Househo
old composiition. As noted, the houssehold compposition grid had no posittive entries in
n 2.5
percent of
o Approach 2C mail screeener returns. An additionnal 1.4 percennt gave respo
onses totalingg 9
adults or more, includ
ding 24 totaling 100 or more. These w
were all consiidered to be ggroup quarteers,
and thus not eligible for
f telephonee follow-up. Review of thhe individual response sugggests that w
while
some of these
t
addressses may in faact be group quarters,
q
othher explanatioons are plaussible, namely (1)
some resp
pondents enttered ages in the boxes in
nstead of num
mber of indivviduals and (22) some enteered a
number representing
r
an estimate of
o the populaation of the nneighborhoood, since the earlier questiions
had asked
d about the neighborhood
n
d.
Household compositiion was also asked in the telephone innterview. Forr those sampled addressess
where bo
oth an Appro
oach 2 mail sccreener and a telephone sscreener weree completed,, we compareed
the total number
n
of ad
dults reporteed. Generallyy, the fewer aadults reporteed in the maiil screener, th
he
more likeely that the saame number was reported
d on the telepphone. If onnly 1 adult waas reported in
n the
mail, 97 percent
p
of tellephone interrviews also reported
r
1 addult; if 2, 88 ppercent agreeed exactly; if 3 or
4, 60 perccent; and if 5 or more, 299 percent. Th
he discrepanccies could be attributable either to
reportingg error or to actual
a
changees in househo
old composittion.
NCVS-CS Pilot
P
Report
5-4
Qu
uestionnaire A
Assessment
5
Telephon
ne Interview
w
The NCV
VS-CS teleph
hone interview
w includes th
hree separatee instrumentss: the househ
hold screenerr,
which is adapted
a
from
m the NCVS Control Card and standaard questionss for an addreess-based sam
mple
to determ
mine whetherr the telephon
ne number iss associated w
with a residennce at the sam
mpled addreess;
the victim
mization screeener, includin
ng most item
ms from the N
NCVS-1; andd the inciden
nt report, or
NCVS-2. Item nonressponse to thee telephone interview
i
inccludes responnses missing because the
respondeent either did not know orr refused to provide
p
an annswer. Unlikke the mail sccreener, theree are
essentiallyy no missingg responses because a queestion was ski
kipped inadveertently.
Househo
old Screenerr. The NCVSS Control Caard questionss included in the Compan
nion Survey P
Pilot
enumeratted adults in the household and obtain
ned basic dem
mographic innformation fo
for each adultt.
The only noteworthy item nonresp
ponse was in
n determiningg the age of hhousehold addults. Month
h and
day of birrth each had 17 percent refused
r
and 6 to 7 percennt unknown; yyear of birth had 14 percent
refused an
nd 4 percentt unknown. If
I the person’s age could nnot be calcullated from daate of birth, the
respondeent was asked
d for an approximate age; of those askked, 21 perceent refused an
nd 5 percentt said
they did not
n know. Th
hose not pro
oviding an approximate agge were askedd to put the person’s age in a
range; agaain, 21 perceent refused an
nd 5 percent did not know
w. Virtually aall of those n
not giving a rrange
acknowleedged whetheer the person
n was 18 or older.
o
Amongg other demoographic item
ms, only high
hest
level of education (2 percent)
p
had more than 1 percent nonnresponse.
Victimizzation Screen
ner (NCVS--1). There waas virtually nno item nonreesponse in th
he NCVS-1
except fo
or household income, whiich had 12 peercent refusaal and 8 perceent unknown
n.
Incidentt Report (NC
CVS-2). Thee incident rep
port includess many questiions, but dep
pending on th
he
type of crrime and oth
her factors, many
m
are skipp
ped in any onne administrration. One set of questio
ons
with relattively high levvels of nonreesponse was characteristi cs of the offfender. For exxample, 30
percent of
o those asked
d did not kno
ow whether there was onne or more thhan one offen
nder, 11 perccent
weren’t suure of the offfender’s age range, 33 peercent didn’t know if the ooffender wass part of a strreet
gang, and
d 38 percent didn’t know whether the offender waas on drugs. IIn addition, 7 percent couuld
not put th
he value of stolen properrty in a price range, and 1 5 percent eitther did not kknow or refuused
the montth of the incid
dent. Of thesse last, all buut 6 percent w
were able to ssay whether the incident
occurred within the 12-month refeerence period
d.
Note thatt the CS hand
dled unknow
wn month of incident diffferently from
m the core NC
CVS. In the ccore
NCVS wiith a 6-month
h reference period,
p
if the respondent doesn’t know
w the month
h the incidentt
NCVS-CS Pilot
P
Report
5-5
Qu
uestionnaire A
Assessment
5
occurred,, the incidentt is discarded
d. For the CSS, the incidennt was retaineed if the mon
nth was not
known orr refused butt the respond
dent confirmed that the inncident occuurred in the p
preceding 12
months.
5.3
Interviewer Debriefing
A debrieffing session was
w held with
h the NCVS CS Pilot inteerviewers, thrree days after completion
n of
data colleection. A totaal of 13 intervviewers atten
nded the sesssion, which w
was facilitatedd by Westat’ss
Telephon
ne Research Center
C
(TRC
C) Operationss Manager whho had traineed the intervviewers. The T
TRC
Project Coordinator
C
who
w managed
d interviewin
ng supervisorrs also particiipated in the debriefing.
Topics in
ncluded gainin
ng cooperatiion, each of the
t pilot instrruments, thee less formal interviewingg style
encouragged during traaining, and reespondent isssues.
Gaining Cooperation
n
Interview
wers reported
d several issuees related to gaining coopperation:
For respo
ondents who
o had already completed th
the mail screeener question
nnaire, manyy felt
they had already comp
pleted the suurvey and woondered why we were callling. Some
respondeents were ann
noyed by the telephone coontact as theey felt they haad already
complied
d with the surrvey request.
Some respondents no
oted that therre was no pooint in complleting the surrvey – that
nothing would
w
come from it, the police
p
don’t ddo anything anyway.
Some inteerviewers no
oted they obtained better cooperation from respon
ndents living in
the city th
han from tho
ose in the sub
burbs.
Househo
old Screener (Control Card
C
Conten t)
The interrviewers noteed several issuues with adm
ministering thhe householdd screener, m
most notably tthat
the conteent was vieweed as very peersonal.
Address, age, race, maarital status, and
a educatioonal attainmeent were askeed about all aadults
listed on the roster an
nd the interviiewers notedd that they haad to use refuusal avoidancce
tactics to get through the screenerr with many rrespondents..
NCVS-CS Pilot
P
Report
5-6
Qu
uestionnaire A
Assessment
5
The first substantive question
q
in th
he householdd screener – “Are your livving quarterss
owned orr being bougght by your orr someone inn your houseehold, rentedd for cash, or
occupied without payyment of cash
h rent?” – prresented som
me difficulties for the
interview
wers.
–
Som
me interview
wers had to exxplain the meeaning of “livving quarterss.”
–
Som
me reported that respond
dents reactedd negatively to being asked this questio
on
(“W
Why do you need
n
to know
w that?” “Thhat’s none of your businesss.”). They had to
exp
plain that BJSS needs to kn
now this in oorder to betteer understandd the nature and
chaaracteristics of
o victims off crime.
Victimiza
ation Screen
ner (NCVS-1
1) Includingg Income/Em
mployment Demograph
hic
Question
ns
There waas not much discussion
d
off the crime sccreener quesstions; the intterviewers feelt those quesstions
were straiightforward and although
h somewhat repetitive
r
theey understoood the rationaale for askingg all
of the crime screener probes.
Regardingg the employyment questio
ons asked at the end of thhe crime screeener (if no rreported
incidents)) or after all incident
i
repo
orts were com
mpleted – thee interviewerrs noted thatt there wasn’tt a
good placce to report employment
e
with nonpro
ofit organizattions or NGO
O’s at item 777 (this readss “Is
your job with
w (1) a priivate compan
ny, business or individuall for wages, ((2) the Federral governmeent,
(3) a Statee, county, or local govern
nment, or (4) yourself, in your own buusiness, profeessional pracctice,
or farm (SSELF-EMPL
LOYED)?”)..
Incident Report (NC
CVS-2)
For respo
ondents with
h a series of in
ncidents, thee interviewerss noted they would like a
more streeamlined wayy to ask abouut things.
Some inteerviewers no
oted that it was unclear whhat to count as a crime, w
when intervieewing
people who
w work in places
p
or professions dealling with the mentally han
ndicapped orr
serially viiolent peoplee (e.g., police, mental heallth professionnals, prison gguards). If a
person is subject to veerbal assault repeatedly inn this type off environmen
nt or with thiis
clientele, does that co
ount? Some respondents ddidn’t think iit should be iincluded, it’s just
part of th
heir job. The interviewerss could use m
more guidance on this, it sshould be
included in the trainin
ng so they can help respoondents reporrt/not reportt as intendedd by
BJS.
NCVS-CS Pilot
P
Report
5-7
Qu
uestionnaire A
Assessment
5
For restarrt interviews – those thatt broke off m
midstream andd are restarteed by a differrent
interview
wer – restartin
ng the surveyy and especiaally completinng the incident summary at
the end of
o the inciden
nt report is very difficult aas the “new”” interviewer has not hearrd
the entiree context for the incidentt.
The interrviewers noteed that they would
w
apprecciate some soort of “bailouut” for mistaaken
reportingg of crime inccidents – thin
ngs that weree mentioned but out of th
he reference
period, or duplicate in
ncidents.
For questtions at whicch the interviewers were iinstructed to answer on th
heir own rath
her
than ask the
t question,, the interviewers were assked if they w
would like to see all caps and
an instrucction (e.g., DID
D R TELL
L YOU THAT
AT SOMETH
HING WAS SSTOLEN
(VERIFY
Y IF NEEDE
ED)?) – theyy all agreed thhat this woulld be helpful..
Other Isssues
Some inteerviewers had difficult interviews, andd were gratefful for the prresence of teaam
leaders (aavailable overr instant messsaging at anyy time) whenn this occurreed.
–
On
ne intervieweer noted when
n she intervi ewed a victim
m, the offendder was sittin
ng
nexxt to the resp
pondent – thiis was a senssitive situation and the intterviewer askked
thee team leaderr to listen in and
a offer anyy guidance w
warranted durring the interrview.
–
Th
he variation in
n content interview to intterview was ““jarring” at tiimes, we hadd few
rep
ports of very serious crim
mes such as doomestic violeence situation
ns, those who
enccountered them mentioneed that they aappreciated tthat their traiining had
inccluded an exaample of exacctly this typee of interview
w.
The interrviewers repo
orted resistan
nce to our reqquest to speaak with multiiple adults peer
householld (in househ
holds with tw
wo or more addults, two orr three were sselected to bee
interview
wed).
At the co
onclusion of the
t initial (ho
ousehold resppondent) intterview, when
n we asked to
o
speak witth others resp
pondents wo
ould ask why we needed tto talk to them, given thaat
they had just completted the surveyy.
–
“W
Why does som
meone else neeed to talk too you?”
–
“Y
You’ll just get the same infformation I j ust told you!!”
For adultt children liviing away at school, there was resistancce to providiing contact
informatiion for them. Parents freqquently woulld not providde a phone n
number to reaach
the selectted young ad
dult.
NCVS-CS Pilot
P
Report
5-8
Qu
uestionnaire A
Assessment
5
Obtainingg contact infformation forr additional aadults when tthe initial aduult was
interview
wed on a personal cell pho
one was difficcult.
–
Th
his included not
n being willling to providde a cell phoone number ffor the other
aduult(s), and alsso being told that the othher adult(s) haad no phonee.
–
Ab
bsent any new
w contact infformation, wee continued tto call the orriginal cell ph
hone
num
mber which often led to refusals
r
by thhe initial/houusehold resp
pondent.
For somee respondentts, different phone
p
numbeers were provvided over tiime – for
example a work phon
ne number may
m have beenn provided by one respon
ndent for ano
other
respondeent, but upon
n calling that we were askeed to not calll the person at work. A b
better
way of geetting to diffeerent phone numbers forr the case woould be helpfu
ful – in this tyype
of situatio
on being ablee to switch back
b
to the hoousehold phoone number would have been
beneficiall.
Sometimes this inform
mation would
d be includedd in messagees left by interviewers
followingg call attemptts, but the system will nott automaticallly dial thosee, it requires uusing
the instruument to swittch to anotheer phone num
mber (and thhis requires ddialing the ph
hone
number currently
c
set for the intervview first, thhen switchingg to another p
phone).
The interrviewers askeed if we could
d add a closinng statementt, for househ
holds with on
nly
one adultt. The curren
nt text states “Let
“
me checck and see iff there are oth
hers I need tto
speak witth” (this is standard for alll cases), theyy wondered iif we could b
bypass that fo
or
single-aduult household
ds and go to a closing/thhank you screeen instead.
Despite the
t difficultiees they experiienced in adm
ministering thhe survey insstruments, th
he intervieweers all
noted thaat they had reeally enjoyed the study – they felt it w
was importantt and useful, and noted th
hat
people would tell them
m a story. Th
hey would likke to work onn the projectt again.
5.4
Behavvior Codin
ng and Monitoring
M
g
In order to
t assess the CATI impleementation of
o CS intervieews, Westat rrecorded all iinterviews wh
here
the respo
ondent gave permission
p
to
o record. Pro
oject staff listtened to manny of the reco
ordings, both
h to
understan
nd how intervviewers and respondents reacted to thhe questionnnaire and as p
part of the typ
peof-crime (TOC) codin
ng process deescribed in th
he next chap ter. Because of the intereest specificallly in
the NCV
VS-1 in the tellephone enviironment, wee selected 100 interviews (of which 977 were codedd), 50
where an incident wass reported an
nd 50 with no
o reported inncident, to haave interview
wer and
respondeent behaviorss coded. Gen
nerally, the idea of behavioor coding is to quantify ddeviations fro
om a
“perfect”” interview byy question an
nd type of deeviation. Wheere such deviiations occurr frequently, there
may be isssues with thee question itsself, with ansswer categoriies provided,, or with inteerviewer train
ning.
NCVS-CS Pilot
P
Report
5-9
Qu
uestionnaire A
Assessment
5
This sectiion will preseent the resultts of the behavior codingg, and includee as well obseervations fro
om
monitorin
ng and codin
ng review of recordings.
r
r
behaviors
b
cod
ded are show
wn in Table 5-1. A code w
was entered fo
or
The interrviewer and respondent
each quesstion from caategories IA, IB, and IIA. Other codee categories w
were used on
nly when the
behavior was observeed. Coders weere encouragged to write nnotes on the coding sheett where prob
blems
occurred describing th
he problem. Four coders received a 44-hour traininng before starrting the reviiew.
Each cod
der listened to
o and coded 23-25 intervviews.
Table 5-1
1.
Interviiewer and respondent be
ehavior codess used in NCV
VS-CS pilot
I. Intterviewer beha
aviors
A. Read qu
uestion exactlly as written
B. Recorde
ed response correctly
c
C. Confirm
med previouslyy reported info
ormation
D. Failed to
t acknowledg
ge previous infformation
E. Repeate
ed all or part of
o the question
F. Used other probe
G. Offered
d other clarification
H. Task-orriented comment
II. Re
espondent beh
haviors
A. Codeab
ble response after
a
question was read once
B. Interrup
pted interview
wer to provide a response
C. Asked for
f clarification
D. Comme
ented on interrview process
E. Volunteered additional information
n
Codes
No change, m
minor change,, major chang
ge
Yes, no, can’tt tell
Occurred
Occurred
Occurred
Occurred
Occurred
Occurred
Codes
Yes, no, can’tt tell
Occurred
Occurred
Occurred
Occurred
A compleete tabulation
n of the resullts is presented in Appenndix C. Theree were no maajor issues wiith
individuaal closed-endeed questions in the NCVS-1 probes. M
Most of the pproblems ideentified weree with
minor changes to the wording of the
t question.. Almost 3,8000 questions were coded;; 16 percent o
of
the questiions were flaagged as mino
or rewordinggs, and aboutt 5 percent ass major rewo
ordings, most
commonlly failing to read
r
through to the end of
o the questioon. None of the other intterviewer
behaviorss occurred in
n more than 2 percent of the questionns overall. Intterrupting the interviewerr with
a responsse (3.3% of questions
q
cod
ded) was the most frequenntly observedd respondent behavior otther
than simp
ply answeringg the questio
on. Respondeents also askeed for clarificcation in 2.4 percent of
questionss coded, and volunteered additional in
nformation inn 2.6 percentt. Thus, mostt interview
sections were
w simple “question,
“
an
nswer, questiion, answer” sequences.
NCVS-CS Pilot
P
Report
5-10
Qu
uestionnaire A
Assessment
5
Appendixx D has a rep
presentation of the questiions that werre reviewed, aannotated wiith findings o
of
note and suggestions for possible revisions. So
ome specific iissues includded:
Respondeents often in
nterrupted thee first crime probe with aan answer beefore the
interview
wer finished reeading the quuestion. The proper interrviewer behavvior is to finiish
reading th
he question, perhaps addiing a processs comment ssuch as “I’m sorry, I’m
required to
t read the entire questio
on.” Most resspondents leaarn quickly th
hat interruptting
will only slow down the interview. But, if the iinterviewer ddoes not finissh the questio
on or
make a prrocess comm
ment, the resp
pondent will likely continnue to interruupt, as the
questionss can seem lo
ong and/or repetitious.
At the en
nd of a group
p of probes, if
i there has bbeen a positivve response tto one or mo
ore,
the intervviewer asks, “What
“
happeened?” and ““How many ttimes did thiss happen?” T
These
questionss do not neceessarily flow with the proobe sequence, and intervieewers strugglled
some gettting the respondents to understand
u
w
what was needded.
Interview
wers were insttructed to record a brief ddescription oof each incident to refer tto
later. Theere was consiiderable variaation in how
w much and w
what informaation interview
wers
wrote in this
t descriptiion, and in th
he probes theey used to eliicit the inforrmation.
The quesstions with th
he most prob
blems identifi
fied, for bothh interviewerss and
respondeents, were thee first crime probe
p
and thhe question oon number off vehicles ow
wned.
Both are long and com
mplex.
Monitorin
ng and codin
ng reviews off the recorded
d interviews yielded the ffollowing observations w
with
regard to the NCVS-22:
The mostt problematicc item was Question
Q
88 ((“Was anythinng stolen?”),, which is
intended to be asked only if it is not
n obvious ffrom previouus questions o
or what the
M
interviewers asked the questionn, or even reaad it as “Was
respondeent has said. Many
anything else stolen?”” Either way, the responsees were thenn often quite misleading to
o
coders.
Generallyy, the flow off the NCVS-2 is not intuiitively obviouus; interview
wers have to b
be
very awarre of what haas been said previously,
p
aand even if thhey perform iimpeccably, some
of the NC
CVS-2 questiions appear to
t be non seqquiturs.
Two concepts criticall to TOC cod
ding, presencce and locatioon, were also
o difficult forr
interview
wers and respo
ondents; inco
orrect interprretation of thhese conceptts led intervieews
down inaappropriate skip trails.
NCVS-CS Pilot
P
Report
5-11
D
Data Processiing
6
a Key Fin
ndings
Chapter Highlights and
Info
ormation on NCVS
N
crime co
oding is not pu
ublicly availablle.
The
e NCVS instrum
ments are diffficult to admin
nister and inacccurate implementation can have a large
e
imp
pact on the typ
pe of crime co
odes.
Wh
hen questions driving skip patterns were recorded
r
inco rrectly by the interviewer, ccrime coding w
was
diffficult. The sam
me was true off interviewer use
u of the “oth
her” category ((which also im
mpacted skip
pattterns).
The
e concept of “p
presence” is not
n well undersstood and diffferences in co
oder opinion ca
aused differen
nces
in coding.
c
The sa
ame was true with
w the definition of a “wea
apon.”
Cod
der reliability was
w low. Almo
ost 20 percentt of incidents were coded differently by the two researrchers
leading the codin
ng process.
Since the primary goaal of the CS Pilot
P
was to assess
a
the feaasibility of thhe study desiggn, postcollection
n data processsing was min
nimal, limited
d to editing vvariables requuired for weiighting and
coding tyype of crime (TOC). New
w variable con
nstruction waas focused onn generating the TOC co
ode
used by BJS
B for estim
mation.
6.1
Crime Classificcation Pro
ocess
Informatiion is not puublicly availab
ble on the TO
OC coding pprocess. Westtat staff was p
provided witth a
flow charrt illustrating how Censuss used the NC
CVS-2 variabbles to determ
mine the app
propriate TO
OC
code. A copy
c
of the fllow chart is provided
p
in Appendix
A
E. Westat also received a taable with the
Census sp
pecifications based on thiis flow chart. This Censuus specificatioons table is p
provided in
Appendixx F. The tablle in Append
dix G summaarizes the Cennsus specificaations; note tthat the codees are
assigned in
i order of severity and so
s may not bee in numericcal order.
Westat ussed these speecifications to
o develop a SAS
S program
m to generatee TOC codes based on th
he CS
data. Thee code was ruun on the raw
w pilot data as
a recorded bby CS intervieewers. In thee course of
working through
t
the TOC
T
codingg step, the ressearch team llearned that a Census codding team revviews
NCVS-CS Pilot
P
Report
6-1
Data
a Processing
6
every inciident recordeed in the NC
CVS. The cod
ders review innterviewer teext and makee a determinaation
of whetheer the raw daata were reco
orded accurattely. Coders w
work from a set of guidellines for upddating
the data, and use a computer-aided
d system devveloped speciifically for thhis purpose. W
When the co
oder is
n, the case is referred
r
to Census
C
Headqquarters (HQ
Q) for additioonal review. C
Census shareed the
uncertain
coder maanual with Westat
W
(provid
ded in Appen
ndix H) and aalso spoke wiith us about some of the HQ
decisions.
Westat sttaff implemen
nted a similar review proccess for the C
CS. Each TO
OC code was reviewed byy the
research team
t
in conccert with the written descrriptions for eeach incidentt. Incidents w
where the texxt
descriptio
on did not match
m
the inteent of the cod
de were idenntified for pottential re-codding. Since th
his
was a feasibility studyy, the team on
nly re-coded the TOC, annd did not reecode the undderlying raw data,
which difffers from Ceensus proced
dures. For example, if a teext descriptioon indicated that a respon
ndent
was “not present” durring a crime, the Census staff would rrecode Q20A
A and Q20B accordingly.
Westat sttaff only reco
oded the TOC
C, and did no
ot edit the unnderlying CSS data.
Based on
n Census documentation and
a discussio
ons with Cennsus staff, keyy factors thatt reviewers
looked fo
or in their asssessments inccluded:
Whetherr the inciden
nt qualifies as
a an NCVSS victimizattion. Some reeports do no
ot
qualify ass a crime. Forr example, a car accident does not quualify as a crim
me (unless th
here
is also an
n apparent inttent to harm)). There are aalso crimes thhat do not qqualify as an
NCVS co
ode-able crim
me. Exampless include bannk fraud, vanndalism, and eexhibitionism
m.
Finally, being
b
a witnesss to a crime (including ggang violencee) was not co
onsidered an
NCVS co
ode-able victiimization.
Whetherr an item waas stolen. Th
here was som
me confusionn associated aas to whetherr
question 88 (somethin
ng stolen) refferred to “soomething elsee” not yet rep
ported. This
seems duue to apparen
nt redundancyy between thhe NCVS-2 aand NCVS-1.. A substantiial
portion of
o Westat’s reevised codes are attributaable to the miiscoding of qquestion 88.
Whetherr stolen item
ms belonged
d to an eligib
ble HH mem
mber. If the stolen propeerty
belonged
d to a minor, then theft co
ould becomee out-of-scoppe, dependingg on the locaation
of the theeft; per Census, theft of property
p
beloonging to a m
minor, such aas toys and
backpackks, are consid
dered out of scope
s
if stoleen away from
m home. There are also caases
where thee respondentt reports thefft of items beelonging to ssomeone outtside the
householld (a friend or
o an employeer, for exampple) – these inncidents werre also considdered
out of sco
ope.
Whetherr a theft insiide the hom
me counts as burglary. T
The review prrocess includded a
check on whether thee offender haad the right too be in the hhome. If the o
offender wass a
guest or repairman
r
(fo
or example) then
t
the thefft was not a bburglary but larceny.
NCVS-CS Pilot
P
Report
6-2
Data
a Processing
6
Whetherr the respondent was prresent durin
ng the victim
mization. Peer Census, thee
respondeent and the offender mustt be in the saame space to count as preesence (i.e.,
presence implies partiicipation in the victimizattion); for exaample, an offfender outsidde in
the yard and
a the respo
ondent asleep
p inside wouuld not be connsidered “prresent.”
Whetherr mentions of
o “vacation
n homes” an
nd “hotel roo
oms” were ttreated
correctlyy. Under som
me circumstan
nces locationns that best fi
fit categories of “vacation
n
homes” or
o “hotel roo
oms” were ap
pplied to an ““other speciffy” category. This coding
would ressult in inapprropriate skip logic and thhe TOC codee often requirred revision.
Whetherr there was a break-in. When
W
the cooded response to question
n 10 did not
match a review
r
of thee summary viictimization ttext, (for exaample a “vacaation home”
coded as “other”), theen questions focused on eevidence of a break-in weere not
c
difficulties w
we experienced were in co
orrectly
administeered. Other classification
classifying garages – again,
a
a review
w of the textt was neededd to correctlyy identify garaage
break-inss. If the text description
d
in
ndicated dam
mage or used a version off the phrase
“broke in
nto the garage,” then the reviewer
r
connsidered it a bbreak-in. Excceptions inclluded
public garages.
Whetherr an object recorded
r
as a weapon iss considered
d a weapon by the NCV
VS.
There aree some weapons reported
d in the NCV
VS-2 that are not considerred to be a
weapon when
w
assignin
ng TOC cod
des. An exam
mple is a Taser gun. If an o
offender sho
oots a
victim wiith a Taser, th
hen it is not considered a weapon by the NCVS. O
On the otherr
hand, if the Taser is used
u
as an ob
bject to hit a pperson, thenn it will be considered a
weapon. The
T strict intterpretation of
o weapons, or when an iitem may be considered a
weapon, resulted in TOC
T
recodes.
Whetherr a thrown object
o
is con
nsidered a w
weapon by th
he NCVS. T
Thrown objeccts
require sp
pecial handlin
ng and review
w. If there w
was serious innjury, then the thrown object
counts ass a weapon. If
I there was minor
m
injury,, then weapoon-status is baased on the aage
of the offfender – if kn
nown to be 12
1 or youngeer, then the thhrown object is NOT
considereed a weapon (if older or unknown,
u
theen it is treateed as a weapo
on). Finally, iif the
respondeent was not in
njured, then Census doess not generallly consider th
he thrown ob
bject
a weapon
n. It may be there
t
are exceeptions (for eexample, largge rocks repeeatedly throw
wn at
the victim
m may be couunted as a weeapon).
Whetherr the inciden
nt qualified as a face-too-face threatt. Threats maade over the
Internet or
o telephone are not conssidered an N
NCVS threat. The same is true of threaats
delivered by a third paarty. Also, th
he research teeam concludeed that threaats (implied o
or
otherwisee) made from
m perpetratorrs in cars wouuld not reachh the thresho
old of a face-toface threaat.
Whetherr a police off
fficer was naamed as the offender. Iff police are n
named as
offenderss then the revviewer must assess whethher the case rrequires re-co
oding. If the
perceived
d offense wass completed in the line off duty, (such as writing a ticket/citatio
on),
then the incident
i
is reemoved. Also
o, mere preseence of a weaapon is not cconsidered in
n
coding th
he crime since most officeers are armedd as part of ttheir job.
NCVS-CS Pilot
P
Report
6-3
Data
a Processing
6.2
6
Whetherr a response of “other/sspecify” affeects the TO
OC code. In ssome cases a
Westat in
nterviewer maay have seleccted “other/sspecify” to ddescribe a feaature of the
incident (for
( example, in describin
ng what was sstolen or thee type of injuury). The
other/speecify responsses were reviiewed to deteermine whethher they coulld be coded tto an
existing category.
c
Notte that this iss the one reviiew process iin which the Westat team
m
recoded the
t variables before sendiing the data tthrough the SAS code to determine th
he
TOC.
Whetherr the inciden
nt was a dup
plicate. Reviiewers checkked incidents reported witthin a
householld to identify duplicates. In
I some casees an incidentt was recordeed more than
n
once in th
he NCVS1; in
i other casess multiple hoousehold mem
mbers reportted the samee theft
(although
h this, interesstingly, was raare).
Sensittivity Ana
alysis
A membeer of the reseearch team reeviewed all in
ncident descrriptions and ddetermined w
whether a revvised
TOC cod
de was warran
nted. To asseess the stabiliity of the codding process, a second m
member of thee
research team
t
revieweed a sample of
o incidents. All personal crimes (n=1173) were sellected for double
coding (th
his includes TOC
T
codes 01
0 through 23).
2 In additioon, a random
m sample (n=
=50) of otherr
coded criimes (TOC codes
c
24 thro
ough 41) was selected, andd a random ssample (n=50) of NCVS--1
reported incidents thaat did not ach
hieve an NCV
VS TOC codde.
Out of th
he 273 cases that
t were double coded, 50
5 were codeed differentlyy by the two reviewers (18%).
Of the 500 discrepant cases,
c
38 werre personal crimes
c
(a 22%
% disagreemeent rate), 3 w
were other codeable crim
mes (a 6% disaagreement raate), and 9 weere from the un-coded inncidents (an 118% disagreeement
rate). Thee first review
wer used the SAS
S data to review
r
the addditional infoormation avaiilable, while the
second reeviewer relied
d solely on th
he text descriiptions. Som
me of the diffeerences weree due to addittional
informatiion available to the first reviewer. Man
ny of the othher differencees were attrib
butable to
differencees in how thee reviewers in
nterpreted th
he presence oof weapons, and also to tthe assessmen
nt of
whether a reported th
hreat counted
d as “face-to--face.”
All differences were adjudicated
a
by four of thee study researrchers: two w
were the prim
mary and the
secondaryy reviewers, and
a two were senior stafff well-versedd in the NCV
VS data and crime statisticcs in
general. Each
E
differen
nce was revieewed and a fiinal code wass determinedd using a com
mbination of the
data, the Census docuumentation and discussion
ns, and expeert judgment by the review
w team.
NCVS-CS Pilot
P
Report
6-4
Data
a Processing
6
Extent of
o Code Reviision. The reesultant revieewed codes w
were then compared with
h the originall
TOC cod
de based on the
t raw data: 22 percent of
o TOC codees were channged in the reeview processs.
This 22 percent
p
(n=3330) was distrributed as folllows:
10.5 perccent is attribuuted to cases initially codeed as out-of-scope but co
oded in-scopee
after the review;
4.2 percent due to casses coded in--scope originnally but codeed out-of-sco
ope after the
review; an
nd
7.1 percent is attributed to changees in the TOC
C code, from
m one in-scop
pe code to a
different code post-reeview.
The majo
ority of cases that were reevised from out-of-scope
o
to in-scope (bullet 1 abo
ove) were chaanged
due to intterviewer and
d respondentt misinterpreetation of queestions 88 annd 89 (theft/attempted th
heft).
As alread
dy indicated, many
m
Westatt interviewers and responndents thoughht that these questions w
were
asking ab
bout somethin
ng “new” or in addition to
t the currennt report (rathher than a co
onfirmation o
of the
theft repo
orted in the NCVS1).
N
Mo
ore than threee-quarters off the cases chhanged from out-of-scope to
in-scope are attributab
ble to the con
nfusion surro
ounding questions 88 (acctual theft) an
nd 89 (attemp
pted
theft). Wee expect mucch of this misinterpretatio
on was due tto interviewer error. The Westat
interview
wers did not administer
a
the NCVS2 veery often, andd based on thhe audio reco
ordings it seeems
that manyy interviewerrs had assum
med that questtions 88 and 89 were atteempting to co
ollect new
informatiion, rather th
han confirmin
ng informatio
on already coollected. Shoould the next phase of thee
Companiion Study utillize an intervviewer-admin
nistered surveey, we wouldd recommendd enhancemeents
to the CA
ATI, to intervviewer trainin
ng, and to intterviewer moonitoring thaat would allevviate this
problem.
Violent Crime.
C
Somee cases were not originallly coded as vi
violent crime but were revvised in the
review prrocess. Theree were 17 inccidents not in
nitially codedd as a violent crime revised to violent ccrime
in the revview process (in this case we define viiolent as incluuding TOC ccodes 01 thro
ough 17). In a
few casess the respond
dent seemed to be reticen
nt to admit thhe crime. Forr example, a rape in which
h the
victim expressed feelings of shamee and self-blaame did not respond affirrmatively to the closed-en
nded
questionss when prom
mpted, but fro
om the open ended descri
ription it is a clear case off rape. In ano
other
case, the respondent described
d
cleear domestic assault but ddid not reporrt an attack in
n the closedended quuestions. Con
nversely, there were 24 cases originallyy coded as vioolent that weere changed
(either to non-violent or to out-off-scope) in th
he review proocess. Many oof these casees involved an
n
affirmativve response to
t the weapo
on question, but
b based onn the text desscription it w
was clear that the
NCVS-CS Pilot
P
Report
6-5
Data
a Processing
6
object waas merely useed to gain enttrance to pro
operty (for exxample, a scrrewdriver or bolt cutters uused
to enter a building or to remove a bike lock).
6.3
Recom
mmendattions
The TOC
C coding process was veryy instructive to the Westaat team and ppoints to chaallenges in ussing
the NCV
VS instrumenttation as the only tool to settle on thee correct TOC
C code. As p
part of the reeview
process th
he Westat research team listened to many
m
intervieew recordingss to determin
ne whether th
he
data captuured accurateely portrayed
d the incidentt. It was cleaar that there w
was sometim
mes a disconn
nect
between the
t descriptio
on of the inccident and th
he individual data points rrecorded. Sin
nce many NC
CVS2
responsess drive skip patterns,
p
a misstep early on
o in the inteerview processs can have m
major
implicatio
ons for the reemainder of the data colleection. For eexample, if a garage was ccoded as “neaar
home” raather than “in
n the home” then questio
ons about eviidence of a bbreak-in are n
never asked. D
Due
to the sen
nsitivity of th
he NCVS2 in
nstrument, wee conclude thhat an unfam
miliar research
h team will fa
face
challengees in administtering the insstrument successfully.
With an interviewer
i
manual
m
more than 900 pagges long, we would anticiipate that eveen Census
interview
wers are challeenged to abso
orb the entirrety of the traaining on a sppecific questtionnaire item
m
when in the
t field. Givven how fraggile the instruument can bee, and given tthe nuance of some key
concepts (such as “prresence” or “weapon”), we
w wonder whhether a postt-interview fo
form could heelp
improve the
t informattion availablee to the Census coders. O
One idea wouuld be someth
hing as simplle as
a checklisst the intervieewer must co
omplete afterr each interviiew, asking ffor more detaails on key
componeents of each incident
i
(for example: loccation, evidennce of a breaak-in, presencce, weapons,,
theft, etc..)
If the nexxt phase of th
he Companio
on Study utilizes telephonne interviewss to collect N
NCVS2 data, we
recommeend the follow
wing changess and enhanccements to thhe design:
Update th
he intervieweer monitoringg plan to incclude both “liive” monitorring as well aas
monitorin
ng of audio recordings.
r
This
T will ensuure we can lissten to each iinterviewer aas
they impllement the NCVS2.
N
(“Livve” monitorinng was unabble to catch m
many interview
wers
as they were
w implemeenting the NC
CVS2.) A pottential challeenge with thiss is that the
monitors would need to be well veersed with N
NCVS and itss nuances.
Incorporate a post-intterview report form that allows intervviewers to an
nswer questio
ons
about keyy details of eaach incident (for examplee, presence, llocation, injuury, etc.)
NCVS-CS Pilot
P
Report
6-6
Data
a Processing
6
Lengthen
n the interviewer training time and speend more tim
me discussingg the relation
nship
between the NCVS1 and
a the NCV
VS2. Also coonsider addinng a discussio
on on TOC
coding to
o the trainingg in order to provide
p
morre context to the interview
wers.
Revise th
he initial studyy introductio
on text in CA
ATI so that reespondents uunderstand th
he
structure of the intervview and the order of queestions. This would includde a descripttion
of the NC
CVS1 (probees to supportt recall and iddentify incideents) and thee NCVS2 (deetails
about eacch incident).
Provide more
m on-screeen help to in
nterviewers ffor critical iteems (e.g., Q10, Q20, Q233,
etc.)
Conduct a post intervview debriefin
ng after the iinterviewer’s first compleeted NCVS2.. Test
derstanding of critical item
ms and coachh them on areeas where diffficulty of
their und
confusion
n was observved.
NCVS-CS Pilot
P
Report
6-7
W
Weightiing
7
a Key Fin
ndings
Chapter Highlights and
We
eights were generated at the
e household, person,
p
and in
ncident level.
The
e NCVS househ
holds weightss are based on
n the weight off a “principal p
person.” The C
Companion Su
urvey
did not have com
mparable scree
ening informa
ation and so in
nstead compu
uted final houssehold weightss
bassed on control totals from th
he Decennial Census.
The
e Companion Survey
S
estima
ates are not bo
ounded, which
h differs from the NCVS which uses the fiirst
wavve panel data to bound the second wave.
The core NCVS has weights
w
at various levels and
a we compputed compaarable weightts for the CS.. The
weights for
f the CS aree, like those for
f the core, at the follow
wing levels: (11) person, (2)) household,, (3)
victimizattion, and (4) incident. Altthough we fo
ollowed the w
weighting proocedures useed in the coree
generally,, for several reasons
r
we used
u
proceduures that weree more approopriate for th
he CS. The
rationale for deviatingg from the co
ore procedures was that nnot all of the core NCVS weighting stteps
are possib
ble or applicaable to the CS.
C For examp
ple, the core is a nationall sample, with
h a rotating p
panel
design an
nd a six montth recall perio
od, but the CS
C is a cross- sectional surrvey with a 122-month recall
conducted only in Ch
hicago.
The majo
or differencess from the co
ore weightingg proceduress are mentionned first, andd then we desscribe
the speciffic procedurees for the CSS in the sectio
ons that folloow. The first--stage ratio eestimate facto
or is
not appliccable for the CS because that is used in
i the NCVSS to adjust foor selecting o
one PSU in nonself-repreesenting stratta; the CS does not have that
t design fe
feature. The bbounding adjjustment for
victimizattions in the first
f wave intterview is nott applied sincce only one iinterview is cconducted fo
or the
CS. The weighting
w
control factor, new permit factor, and w
weighting facctor cap are aalso not
applicablee. The within
n household noninterview
w factor usedd in the NCV
VS to adjust ffor
nonrespo
onding person
ns within ressponding houuseholds is a weighting cllass adjustmeent; we chosee to
use rakingg to control totals for this purpose. The
T core houssehold noninnterview factor could nott be
applied because in thee CS we do not
n identify a head of houusehold for nnonrespondin
ng householdds.
Another important
i
diifference is th
hat the core NCVS
N
uses tthe “principaal person” prrocedure to
compute a household
d weight, wheereas the CS uses the inveerse of the hoousehold’s in
nclusion
NCVS-CS Pilot
P
Report
7-1
Weighting
7
a
frame. The CS ho
ousehold weiights are thenn adjusted fo
or screener
probabilitty from the address
nonrespo
onse and posttstratified to Census 20100 household totals.
The CS uses
u the collection year fo
ormat for all analyses,
a
wheere all victim
mizations are counted in th
he
year repo
orted (as oppo
osed to the data
d year form
mat, where vvictimizationss are countedd in the year they
occurred)). This is con
nsistent with the data colleection proceedure used inn the CS. With
h the collectiion
year form
mat, the perso
on crime ratee is the sum of
o all victimizzation weights for 12 months dividedd by
the sum of
o all person weights for 12 months. In
I the core, tthe rate is muultiplied by 2 to compenssate
for the 6--month recall period; sincce the CS askks about crim
mes in the last 12 months,, the factor o
of 2 is
not needeed.
Each step
p of weightin
ng was done separately fo
or the Approaach 1 and Appproach 2 samples. This w
was
done because the goall is to compaare estimates from these ssamples sepaarately to thee core NCVS. The
Approach
h 1 and Apprroach 2 samp
ples are analyyzed separateely, and theree are no plan
ns to combinee the
two.
7.1
House
ehold Weights
7.1.1
Househ
hold Base Weights
The ABS frame was stratified
s
into
o two strata: (1)
( City of Chhicago, and ((2) Remaindeer of Chicago
o
CBSA an
nd a stratified
d simple rand
dom sample of
o addresses w
was selectedd. The City off Chicago strratum
was samp
pled at a rate that was 1.5 times that used in the rem
ple of 12,5000
mainder straatum. A samp
addressess was selected
d for Approaach 1 and 14,,000 addressees for Approoach 2. The h
household baase
weight is equal to the inverse of th
he probabilityy of selectionn for the houusehold’s adddress. As a reesult,
the samplled househollds for each approach
a
weeight up sepaarately to the entire Chicaago CBSA.
In Appro
oach 2, 50 percent of the households that
t did not rrespond to thhe mail screeener (by a
specified date) but had a telephone number avvailable were subsampled.. Only the suubsampled m
mail
nonrespo
ondents weree sent for the household telephone
t
intterview. Thiss subsamplingg is reflectedd in
the houseehold base weight
w
for thee household telephone
t
scrreener for thhe Approach 2 cases but n
not
for Appro
oach 1 wheree there was no
n subsampliing. The houusehold base weight for th
he mail screeener
does not include this subsamplingg.
NCVS-CS Pilot
P
Report
7-2
Weighting
7.1.2
7
Househ
hold Nonre
esponse Adjustmen
A
nts
In the Ap
pproach 1 sam
mple, a houssehold is considered to bee a respondennt if there is a completedd
householld NCVS screeener instrum
ment on the telephone.
t
A nonresponsse adjustmen
nt was done ffor
the houseehold telepho
one screener so that the households
h
w
with a completed screenerr represent n
not
only the nonsampled
n
households, but also the nonrespondding househoolds. Weightin
ng classes for the
nonrespo
onse adjustmeent were defi
fined by whetther or not thhe householdd’s address h
had a matchin
ng
telephonee number fro
om the vendo
or, by metro status (city oof Chicago, rremainder off Cook Countty,
other couunties in Chiccago CBSA) and race chaaracteristics oof the Censuss 2010 tract w
where the
householld is located (<=30%
(
Black and Hispaanic populatiion, >30% B
Black and Hisspanic
populatio
on).
In the Ap
pproach 2 sam
mple, becausse of the two phases in thhe householdd screener (m
mail and
telephonee), a househo
old may be a respondent at
a each level.. A householld is a respon
ndent at phasse 1 if
it returns the completted mail quesstionnaire. A subsampled household aat phase 2 is a respondentt if
there is a completed household
h
NCVS screeneer instrumentt on the telepphone. The ddefinition of
phase 2 household
h
ressponse in Ap
pproach 2 is the same as tthe definitionn of househo
old response in
the Approach 1 samp
ple.
A separatte nonrespon
nse adjustmen
nt was done for both thee mail and tellephone screeeners, using the
same weighting classees as in Approach 1 abovee. The nonreesponse adjusstment factor for the
telephonee screener waas applied to the househo
old base weigght that refleccts the subsaampling of m
mail
screener nonresponde
n
ents prior to administerin
ng the telephoone screenerr in Approach
h 2. Thus theere
are two household
h
weeights for Ap
pproach 2, co
orrespondingg to the mail and telephon
ne screeners.
Approach
h 1 only has one househo
old weight beecause there w
was no subsaampling.
The nonrresponse adjuustment facto
or for the maail and telephhone screeneers was calcullated within
weightingg classes as:
NR
R Adj Fac
ieligible
hh basse wt i .87 *
hh base wt
ieligible responddent
NCVS-CS Pilot
P
Report
hh base wt
iunknown residenntial status
7-3
i
i
,
Weighting
7
where thee numerator includes an estimate
e
of th
he number oof residential addresses am
mong those
whose residential statuus is unknow
wn. The perceent residentiaal was estimaated to be 877 percent usin
ng
the CASR
RO approach
h (CASRO 19982).
7.1.3
Househ
hold Postsstratificatio
on
The nonrresponse-adjuusted househ
hold weights for the two aapproaches w
were poststraatified separaately
to Census 2010 occup
pied househo
old control to
otals based onn geographicc areas and teenure. The
weightingg classes weree constructed
d using a cro
oss-classificatition of geogrraphic area byy tenure (own
neroccupied or renter-occcupied). Thee geographic areas were ddefined as folllows:
Three sub
bdivisions off the City of Chicago (low
w, medium, hhigh), based o
on crime ratees
obtained from Chicaggo Police Dep
partment datta (gis.chicaggopolice.org, retrieved
9/6/12).
Remaindeer of Cook County,
C
Illino
ois, excludingg the City of Chicago
The other 13 countiess in the Chicago CBSA.
The map below show
ws the three suubdivision of the City off Chicago witthin Cook Co
ounty, based on
their prop
perty and vio
olent crime raates shown in
n the plot.
NCVS-CS Pilot
P
Report
7-4
Weighting
Figure 7-1
1.
7
Crime rates from Chicago
C
Policce Departmeent
Three sub
bdivisions off the City of Chicago (low
w, medium, hhigh), based o
on crime ratees
obtained from Chicaggo Police Dep
partment datta (gis.chicaggopolice.org, retrieved
9/6/12).
Remaindeer of Cook County,
C
Illino
ois, excludingg the City of Chicago
The other 13 countiess in the Chicago CBSA.
The final household weights
w
sum to 3,475,726 for each appproach.
7.2
Person
n Weightts
The perso
on weights were
w computeed in two steeps. First, thee base weightt for the sam
mpled person was
computed
d and multiplied by the household
h
weeight. Secondd, the weight was then rakked to account
for nonreesponse and undercoveraage.
NCVS-CS Pilot
P
Report
7-5
Weighting
7.2.1
7
Person Base Weiights
The withiin-household
d person basee weight wass calculated aas the ratio off the numberr of adults in
n the
householld to the num
mber of adultts selected fro
om the houseehold. The ooverall person
n base weigh
ht was
calculated
d as final houusehold weigght times the within-houseehold personn base weight. For Appro
oach
2, the finaal household
d weight for the
t telephonee screener waas used to caalculate the p
person weightt.
7.2.2
Person Weight Raking
R
The perso
on base weigghts were rakked separatelyy for each appproach to thhe same popuulation control
totals bassed on age, seex, race, and geographic area
a from thee 2010 censuus for the Ch
hicago CBSA. The
raking haad three dimeensions:
Geograp
phic areas ass defined ab
bove.
Age by race.
r
For race, the three categories
c
weere: Hispanicc, black non-H
Hispanic, an
nd
other non
n-Hispanic. The
T age classses were: 18 tto 24, 25 to 334, 35 to 44, 45 to 54, 55 to
64, 65 or more.
Age by sex.
s The sam
me age classess used in dim
mension 2 werre used.
Within th
he age 18 to 24
2 category, race
r categoriies were collaapsed for thee second dim
mension and
gender was
w collapsed for the third
d dimension. This was donne so that a m
minimum off 40 respondeents
were avaiilable in the age
a category to prevent exxcessively larrge raking addjustment facctors.
7.2.3
Person Weight Trrimming
The distriibution of th
he raked person weights was
w skewed tto the right ddue to some vvery large ouutlier
weights. These
T
large weights
w
resulted primarilyy from the suubsampling oof adults in laarge househo
olds
and low response
r
ratee in 18 to 24 age group (eespecially for Blacks and H
Hispanics). T
To reduce thee
potential problems associated with
h outlier weigghts, the rakeed weights w
were trimmedd.
Trimmingg was done separately
s
by approach, an
nd the trimm
med weights w
were raked to
o the same
control to
otals so that the
t control totals
t
were prreserved. Thee initial trimm
ming thresho
old was set to
o the
99th perccentile of the person distrribution for each
e
approacch. The trimm
ming and rakking was repeeated
NCVS-CS Pilot
P
Report
7-6
Weighting
7
he result was that 1.7 perccent of the w
weights were
iterativelyy as needed to avoid largee weights. Th
trimmed for Approacch 1, and 2.4 percent weree trimmed foor Approach 2.
The final person weigghts for each approach suum to 7,083,2295.
7.3
Victimization and
a Incide
ent Weigghts
The victim
mization weiight was calcuulated as the person weigght for personal crimes, aand the houseehold
weight fo
or property crrimes. The only differencce from the ccore NCVS iss that the bounding
adjustmen
nt is not used
d, since each
h household is
i included onnly once in th
the sample raather than
through a panel.
The incid
dent weight was
w also calcuulated the sam
me as in the core NCVS, as the final p
person weigh
ht
times the multiple victim adjustmeent factor, wh
here the mulltiple victim aadjustment ffactor is the
inverse of the numberr of victims in
i the inciden
nt (the numbber of victimss was truncatted at 10).
The victim
mization weiights sum to 2,702,943 an
nd the incidennt weights suum to 2,603,4429 for inciddents
that weree classified as in-scope forr the NCVS.
Table 7-11 is a general guide to how
w the househ
hold, person, victimization and inciden
nt weights arre
used. Exaamples of ho
ow to use thee weights to calculate
c
victiimization rattes and otherr estimates arre
given in Appendix
A
I.
Table 7-1
1.
Generral guide to using
u
the hou
usehold, persson, victimiza
ation and inccident weightts
Weight
Household
d Weight
Person We
eight
Victimization Weight
Incident Weight
W
7.4
Usse for Calcula
ating:
Household prope
erty crime ratees, household
d level estimattes
Person violent crrime rates, peerson level esttimates
Household prope
erty or person violent crimee rates, victimiization level
esttimates
Inccident level estimates
Varian
nce Estim
mation
Standard errors of thee estimates will
w be compuuted using a jjackknife repplication meth
hod. The
replicatio
on process rep
peats each sttage of estimation separattely for each replicate. Th
he replication
n
NCVS-CS Pilot
P
Report
7-7
Weighting
7
u
for obttaining stand
dard errors foor complex sttatistics. Thee standard errrors
method is especially useful
omputed usin
ng the comp
plex survey daata analysis ppackage WesV
Var, or otherr software
may be co
packages that use repllication meth
hods such as Stata, SAS, SSUDAAN, oor R. Variablees are also
ho wish to use Taylor series linearizatioon to compuute standard errors.
included for users wh
Because the
t sample was
w a simple stratified
s
rand
dom sample of addressess, we used a p
paired jackkn
nife
replicatio
on method, aggain separateely for each approach.
a
Thhe addresses w
within stratuum were first
paired in the order sam
mpled and raandomly assigned to be eeither unit 1 oor unit 2. The pairs then w
were
assigned to
t variance stratum sequeentially starting with 1 annd going to 800 and then reestarting at 1 and
continuin
ng. As a resullt, for each ap
pproach therre are 80 variiance strata aand each reco
ord has a variiance
unit of either 1 or 2. The
T variable for the stratuum is called V
VARSTRAT
T and for the unit is calledd
VARUNIIT. Once thee replicate vaariables were defined in thhis manner, tthe exact sam
me set of
weightingg procedures was done seeparately for each replicatte.
A set of 80
8 jackknife replicate
r
weigghts was creaated for eachh household, person, and incident.
Standard errors can be computed using WesVaar by includinng the approopriate set off replicate weeights
(househo
old, person, victimization
v
or incident) and using thhe JK2 replicaation methodd (see Rust aand
Rao (1996)). This sam
me method caan be implem
mented in othher replicatioon packages b
by using a facctor
of unity to
t multiply th
he squared diifferences off the replicatee estimates frrom the full ssample estim
mate,
i.e.,
G
v(ˆ) c (ˆ( g ) ˆ) 2
g 1
mate of the fuull-sample estimate baased on the oobservations included in tthe
where ˆ( g ) is the estim
g-th replicate, c = 1, and
a G is the total numberr of replicatees (80 in this case).
For Taylo
or series proccedures a witth-replacemeent approxim
mation will roughly producce the same
standard errors, althouugh these lin
nearization prrocedures doo not fully account for thee
poststratiification and raking adjustments. For example, in tthe SAS survvey procedurres VARSTRA
AT is
the “Nestt” variable an
nd VARUNIIT is the “Cluuster” variabble.
NCVS-CS Pilot
P
Report
7-8
Weighting
Referencces
Council of
o American Survey Reseaarch Organizzations (CASSRO). (1982). On the defi
finition of
response rates (special
(
reporrt of the CASSRO task forrce on comppletion rates, Lester R.
Fraankel, Chair).. Port Jeffersson, NY.
Rust, K.F
F., and Rao, J.N.K
J
(1996). Variance esstimation forr complex surveys using rreplication
techniques. Stattistical Methodds in Medical Research,
R
5, 2881–310.
NCVS-CS Pilot
P
Report
7-9
7
Esttimatio
on Resu
ults
8
a Key Fin
ndings
Chapter Highlights and
The
e Companion Survey
S
producced higher victtimization ratees than the NC
CVS.
The
e higher victim
mization rates for the Compa
anion Survey ooccur even though it has a much lower
response rate forr population groups
g
that typ
pically experieence more victtimizations.
The
e mail screene
er used in App
proach 2 was a good predicttor of whetherr a victimizatio
on would be
rep
ported in the te
elephone interview.
The
ere was no evidence of interrviewer effects in reporting victimizationss.
8.1
Victimization Rates:
R
Co
ompariso
on with National N
NCVS
Estima
ates and Estimate
es from tthe Chica
ago Policce
Departtment
In this section we preesent estimateed victimizattion rates fro m the two appproaches taaken in the CS,
along witth the nationaal NCVS estiimates for 20011 and estim
mates of crim
mes reported to the Chicaago
Police Deepartment. Section 8.1.1 discusses
d
pottential sourcees of differennces between
n the CS, thee
NCVS, an
nd the Chicaago Police Deepartment reecords. Sectioon 8.1.2 presents estimatees for properrty
victimizattions, and Seection 8.1.3 presents
p
estim
mates for viollent victimizations.
8.1.1
Why Miight Estim
mates from
m the NCVS
S, the CS, and the C
City of Chiccago
be Expe
ected to Differ?
D
Before co
ollecting dataa, it was conjeectured that estimates froom the CS m
might differ frrom the NCV
VS
estimates. The CS dessign has a number of diveergences from
m the NCVSS design whicch might cause
the estimates to differr. We review some of the differences below.
Target population.
p
The
T NCVS target
t
populaation consistss of persons aged 12 and older
who resid
de in households or group
p quarters; ppersons who are living in iinstitutions ssuch
as prisons and nursingg homes or are
a living in m
military barraacks are excluuded. The CSS
target pop
pulation con
nsists of perso
ons aged 18 aand over whho reside in h
households. T
The
NCVS-CS Pilot
P
Report
8-1
Estimattion Results
8
CS samplling frame do
oes not incluude addressess of group quuarters, so peersons such aas
dormitoryy residents arre included in
n the CS onlly if they are included on a householdd
roster forr a household
d in the ABSS frame.
4
Coverage. The list frrame used byy the CS has aan estimated coverage off 97 percent o
of
o nontelephon
ne
residentiaal households. All intervieews were connducted by teelephone, so
househollds were excluded from th
he survey. Thhe frame wass intended to
o cover only tthe
residentiaal population
n, and group quarters (9 oor more unreelated adults) were to be
excluded.. We anticipaated that the ABS frame w
would includde certain gro
oup quarters, such
as Assisteed Living cen
nters where the
t address iss unit-based rrather than ccenter-based.. If a
mail screeener was retuurned indicatting 9 or morre adult residdents, the adddress was
considereed out of sco
ope4.
Sponsorsship. The NCVS data aree collected byy the Censuss Bureau, wh
hile the CS daata
were collected by Westat under th
he sponsorshiip of the Burreau of Justicce Statistics.
Subsamp
pling. The NCVS
N
attemp
pts to interview all personns aged 12 an
nd over in
selected households.
h
The
T CS interrviewed two rrandomly sellected person
ns age 18 or over
in the selected househ
holds. Subsam
mpling adultss and excludi
ding minors w
would be exp
pected
to result in
i fewer repo
orted propertty crimes.
Boundin
ng and recalll period. Th
he NCVS askks about victiimizations occcurring in th
he
last 6 months, while the
t CS asks about
a
victimizzations occuurring in the llast 12 month
hs.
Moreover, the second
d through sevventh NCVSS interviews aare often bouunded by thee
previous interview so that telescop
ped victimizaations are noot reported fo
or those
interview
ws. NCVS resspondents on
n later interviiews may alsoo have panel conditioningg
effects ass they becom
me more familliar with the survey. The lack of a bouunding intervview
in the CSS means that telescoping may
m occur. A
Additionally, respondents may have beetter
recall for more recentt incidents th
han for incideents occurrinng 11 or 12 m
months ago.
Interview
wer and mode effects. Previous
P
studdies with the NCVS havee shown that
interview
wer experience is related to
o numbers oof victimizatioon reports reeceived. In 20006,
with the introduction
i
of new prim
mary samplingg units (PSUs) and the traansition to C
CAPI,
estimated
d victimizatio
on rates increeased, and CA
API interview
ws in the inco
oming PSUss had
more victtimization reports than CAPI
C
intervieews in the continuing PSU
Us (Rand, 20008).
This effecct was thougght to be due to new interrviewers. Truuman and Plaanty (2012)
discuss exxperiments conducted
c
duuring 2011 too investigate tthe effects of field
representtative refresh
her training. Since
S
all of thhe intervieweers in CS werre new to thee
survey, siimilar increasses in victimiization rates m
might be exppected. Centrralized CATII was
used for the
t CS, with no initial in--person contaact as with thhe NCVS, an
nd the mode may
lead to diifferent respo
onses. Rand (2009) suggeests that the aamount of tim
me spent on the
screener questionnair
q
e (NCVS-1) may decreas e victimizatioon reports, n
noting that fo
or the
NCVS: “A
Analyses of time
t
stamps (available sinnce the surveey was autom
mated in 20066)
We believe that
t some househo
olds may have enterred ages in the houusehold compositioon matrix, rather thaan the count of inddividuals of a certaiin
gender-age group.
g
Since we co
ould not determinee which of these surrveys were indeed ggroup quarters andd which were measuurement error we ddecided
to treat all su
uch returns as grou
up quarters so longg as the count of ad
dults exceeded 8.
NCVS-CS Pilot
P
Report
8-2
Estimattion Results
8
indicate that
t many intterviewers are routinely nnot taking suffficient time to ask the fuull
battery off questions.””
Coding. As described
d in Chapter 6 the type-oof-crime codiing for NCVS is a compleex
process. The
T CS codeers, new to th
he process, m
may classify ccrimes differeently than NC
CVS
coders. In
n addition, th
he NCVS disscards victim
mization reporrts that are m
missing the m
month
in which the victimizaation occurreed, while the CS includes any reports that the
respondeent said occurrred within th
he past 12 m
months.
Nonresp
ponse. A prio
ori, we expeccted differingg nonresponsse rates to bee the biggest
potential source of differences bettween the NC
CVS and CS. The divergeence could h
have
gone in either
e
directio
on. The higheest victimizattion rates aree typically fouund among
demograp
phic groups that
t tend to have
h
the low
west responsee rates, so it m
might be exp
pected
that nonrresponse wouuld decrease reported victimizations.
On the other hand, viictims may be
b more likelyy to respond to the surveey because they
are more interested in
n the topic. The
T study by Duhart et al. (2001) supp
ports this vieew,
finding th
hat NCVS reespondents in
n hard-to-conntact or initiaal refusal houuseholds repo
orted
slightly feewer propertty victimizatio
ons. A test w
with the Scotttish Crime an
nd Victimisattion
Survey (H
Hope, 2005) found lower victimizationn rates amonng those interrviewed in peerson
(67% resp
ponse rate) th
han among those
t
intervieewed by telepphone (49% estimated
response rate); telephone refusers who were laater converteed had lower victimization
n
rates than
n the initial teelephone respondents. Al
Although the nnon-contact rate was sligghtly
lower forr the telephon
ne survey (9%
%) than the iin-person surrvey (13%), tthe refusal raate
was 41 peercent for the telephone survey
s
but onnly 17 percennt for the in--person surveey.
Using thee design weigghts, Hope (22005) found the followingg victimizatio
on rates in th
he
survey:
Household
d crime
Personal crime
c
Telephone,
all
15.5
7.2
In-perso
on
14.1
1
4.0
0
TTelephone,
convverted refusalss
13.1
4.6
Hope (20005) also foun
nd that usingg weighting aadjustments bbased on dem
mographics ddid
not resolvve the differeences in victiimization rattes for the tw
wo surveys.
Data werre available frrom the City of Chicago police
p
departtment on num
mber of crim
mes for each ttype
that occuurred within each
e
Chicago
o Communityy Area duringg the precediing 12 month
hs. Data werre
retrieved from http://
/gis.chicagop
police.org on
n September 6, 2012, for tthe 12-montth period of
Septembeer 6, 2011 thrrough Septem
mber 5, 20122. The Chicaggo reports innclude some ccrimes that aare
not classiified as crimees in the NCV
VS. They also
o include crim
mes that are out-of-scope for NCVS,, such
as arson, homicide, an
nd crimes thaat occur to minors
m
or to ccommercial eestablishmen
nts. The largeest
potential sources of expected diffeerences betw
ween the Chiccago statisticss and those ffrom the
Companiion Survey orr the core NC
CVS are: (1) the NCVS reecords victim
mizations at tthe respondeent’s
residencee, while the police
p
recordss list victimizzations at thee place of occcurrence or rreporting. Th
his
NCVS-CS Pilot
P
Report
8-3
Estimattion Results
8
h high numbeers of person
ns commutingg in, such as the Chicago Loop, will h
have
means that areas with
m the police records thann in NCVS, aand suburban
n areas will h
have
relatively higher crimee reports from
T Chicago
o Police Depaartment recoords contain only crimes tthat
relatively lower crime reports. (2) The
olice. (3) Thee Chicago Police Departm
ment defines types of crim
me differentlyy
are reported to the po
N
Violeent crime, for example, in
ncludes homiicide and maanslaughter an
nd property
than the NCVS.
crime inccludes arson and
a vandalism
m.
The Chiccago Police Department
D
records,
r
thereefore, would not be expected to accorrd with estim
mates
from eith
her the NCVSS or the CS. However, th
hey would be expected to be correlated with the
victimizattion rates forr the geograp
phic areas.
8.1.2
Propertty Victimizzation
Table 8-11 gives the estimated prop
perty victimizzation rates ffor Approachh 1 and Apprroach 2, alon
ng
with 95 percent
p
confidence intervaals. The prop
perty victimizzation rates w
were calculatted using the final
householld weights deescribed in Seection 7.1. Th
he replicate hhousehold w
weights were uused to calcuulate
the confid
dence intervaals. The 20111 national NC
CVS rates foor crimes are from Trumaan and Plantyy
(2012). Note
N that the scope differss for the estim
mates from tthe CS and thhe national N
NCVS: the C
CS
estimates are for the adult
a
populattion of the Chicago
C
CBSA
A, while the nnational NCV
VS estimatess are
for the naational popullation aged 12 and higherr. For that reeason the lastt column of T
Table 8-1 givves
NCVS esstimates calcuulated using the
t respondeents aged 18 and higher inn the Chicaggo CBSA for the
3-year period 2009-20011. These esstimates havee larger standdard error thaan the nation
nal NCVS
C
and resstricted to thhe adult popuulation.
estimates, but are speccific to the CBSA
d property viictimization rrates from thhe CS are app
proximately ttwice
Table 8-11 shows that the estimated
those of the
t National NCVS. The rates for Ap
pproach 1 andd Approach 2 are not siggnificantly
different. Within the city
c of Chicago, the trend
d in victimizaation rates foollows the sam
me general
direction as the record
ds from the Chicago
C
Poliice Departmeent. The ratioos of the aveerage of the
Approach
h 1 and Apprroach 2 victim
mization ratees to the Chiccago Police D
Department rates are
approxim
mately 6.5 forr the low-crim
me poststratuum, 4 for thee medium-crim
me poststrattum, and 1.5 for
the high-ccrime poststrratum. We co
onjecture thaat may be beccause the areeas in the “hiigh crime”
poststratuum include areas
a
with heaavy commuteer influx suchh as the Chiccago Loop, aand some of tthe
crimes recorded by th
he Chicago Police Department in thosse areas actuaally occurredd to commuteers
who resid
de in one of the
t other areeas. This conjjecture is suppported by exxamining thee ratios of
victimizattion rates in demographicc and geographic domainns for Approaach 1 and Ap
pproach 2 to the
NCVS-CS Pilot
P
Report
8-4
Estimattion Results
8
bles
three-year NCVS victtimization rattes in the Chicago CBSA,, given in thee last two collumns of Tab
8 In Table 8-2, the ratio
os of the CS to Chicago N
NCVS victim
mization ratess are somewh
hat
8-1 and 8-2.
larger forr the low-crim
me poststratuum than for the
t other fouur geographicc poststrata, but not to th
he
same exteent as the rattios with the Chicago Police Departm
ment rates.
The ratios of CS to Chicago NCV
VS victimizatiion rates in thhe last two columns of T
Tables 8-1 andd 8-2
are impreecise becausee of the wide confidence intervals
i
of aall the estimaates. With thaat caveat, tho
ough,
the ratioss (Approach 2 rate)/( Chiicago NCVS rate) appear to exhibit leess variabilityy across
geograph
hic domains than
t
the ratio
os (Approach
h 1 rate)/( Chhicago NCVSS rate), altho
ough the
differencee in variances is not signiificant.
Note thatt Approach 1 appears to follow a diffferent patternn for the low
w, medium, an
nd high crime
areas than
n Approach 2 and the esttimates from
m the Chicagoo Police Depaartment. Thee wide confiddence
intervals for the propeerty victimizaation rates frrom both Appproach 1 andd Approach 2 indicate th
hat
the rankin
ngs for the th
hree types off crime areas are not signiificantly diffeerent for App
proach 1 andd
Approach
h 2. Note thaat the ratios of
o the Appro
oach 1 and A
Approach 2 raates to those of the Chicaago
Police Deepartment fo
ollow the sam
me basic patteern: the ratioo is highest foor the “low ccrime” areas aand
lowest for the “high crime”
c
areas. This divergeence may occcur because tthe CS recordds victimizattions
accordingg to the victim
m’s residencee while the Chicago
C
Policce Department records viictimizations at
the place where the viictimization occurred.
o
Th
hus, thefts thhat occur to a commuter ffrom a “low
crime” arrea while at work
w
in a “higgh crime” areea are recordded in the “hhigh crime” area by the
Chicago Police
P
Deparrtment but in
n the “low crime” area in the CS.
NCVS-CS Pilot
P
Report
8-5
NCVS-CS Pilot Report
TTable 8-1.
Estimated
E
property victimization rates, per thoussand householdss; 95 percent co nfidence interva
als are given belo
ow the
estimates.
e
The la
ast two columns give the ratios of
o the CS victimization rates to tthe NCVS victimiization rates from
m the
Chicago
C
CBSA.
A
All property crime
B
Burglary
M
Motor vehicle thefft
TTheft
H
HU owned
H
HU rented or other
8-6
LLocal questionnaire
G
Generic questionn
naire
CS Appro
oach 1
273
3
[241, 304]
3
49
9
[36,6
63]
3
[0, 6]
6
207
7
[179, 235]
2
265
5
[224, 305]
3
288
8
[227, 350]
3
CS Ap
pproach 2
287
2
[252
2, 321]
56
[4
42,71]
9
[4
4, 14]
196
1
[171, 222]
269
2
[225
5, 312]
321
3
[246
6, 396]
307
3
[247
7, 367]
266
2
[226
6, 306]
Nattional NCVS,
2011
138.7
29.4
5.1
104.2
NCVS,
Chicago CBSA,
C
2009-2011
119
[103, 136]
28
[20, 36]
6
[3, 9]
86
[74, 97]
105
[85, 125]
148
[118, 178]
Ratio of CS
S
Approach 1 to N
NCVS,
Chicago CBS
SA
2.3
Ratiio of CS
Approach
h 2 to NCVS,
Chica
ago CBSA
2.4
1.8
2.0
0.5
1.5
2.4
2.3
2.5
2.6
1.9
2.2
Estimation
Results
E
i
i R
l
8
NCVS-CS Pilot Report
TTable 8-2.
Estimated
E
property victimization rates, per thoussand householdss, for geographicc poststrata; 95 percent confidence
in
ntervals are give
en below the estimates
A
All property crime
C
City of Chicago
Chicago, low crim
me
Chicago, medium
m
crime
Chicago, high crime
8-7
R
Remainder of Coo
ok
C
County
R
Remainder of CBS
SA
CS Appro
oach 1
273
3
[241, 304]
3
391
1
[325, 457]
4
387
7
[293, 480]
4
418
8
[297, 539]
5
373
3
[261, 485]
4
252
2
[184, 320]
3
204
4
[156, 251]
2
CS Ap
pproach 2
287
2
[252
2, 321]
416
4
[343
3, 489]
306
3
[208
8, 404]
466
4
[336
6, 596]
516
5
[371, 661]
247
2
[169
9, 326]
221
2
[173
3, 270]
Chicago
Po
olice Dept.
109
66
109
171
NCVS, Chicago
CBSA,
2009-2011
119
[103, 136]
156
[116, 195]
93
[54, 133]
167
[92, 243]
239
[150, 327]
106
[77, 135]
105
[84, 126]
Ratio of CS
S
Approach 1 to N
NCVS,
Chicago CBS
SA
2.3
Ratiio of CS
Approach
h 2 to NCVS,
Chica
ago CBSA
2.4
2.5
2.7
4.2
3.3
2.5
2.8
1.6
2.2
2.4
2.3
1.9
2.1
Estimation
Results
E
i
i R
l
8
Estimattion Results
8.1.3
8
Violent Victimization
Tables 8--3 to 8-5 givee estimates fo
or violent vicctimization raates from thee CS and NC
CVS. The viollent
victimizattion rates weere calculated
d using the fin
nal person w
weights descriibed in Sectio
on 7.2, and tthe
replicate weights
w
weree used to calcculate the con
nfidence inteervals. As witth property ccrime, the
estimated
d victimizatio
on rates for violent
v
crime are higher foor both apprroaches of th
he CS than fo
or the
national 2011
2
NCVS or the 2009-22011 NCVS in the Chicaago CBSA. Thhe victimizattion rate for
Approach
h 1 is higher than that forr Approach 2,
2 although thhe differencee is not signifficant (p-valuue =
.09). The rates for serious violent crime are siggnificantly diffferent for thhe two appro
oaches, with p
pvalue = .002.
Table 8-55 gives a poteential explanaation for the discrepancy in violent viictimization rrates for the two
approach
hes. The estim
mated rates fo
or men are ap
pproximatelyy equal for A
Approach 1 aand Approach
h 2,
but the raate for womeen is 100 in Approach
A
1 and
a 28 in Ap proach 2. Paart of that diffference is duue to
women who
w report multiple
m
victim
mizations in Approach
A
1 (Table 8-7 shhows that 699 per thousan
nd
women reeport at leastt one violent victimization
n in Approacch 1, comparred with 26 p
per thousandd
women in
n Approach 2).
2 It is uncleear why wom
men would reeport differennt violent vicctimization leevels
in Approach 1 and Ap
pproach 2, allthough this difference apppears to be centered in h
households w
with
more than
n one adult. With the info
formation avaailable, we arre able only tto speculate aabout the possible
cause, if it
i is not mereely a statisticaal anomaly.
As was ob
bserved for property
p
crim
me, both App
proach 1 andd Approach 2 produce higgher estimatees of
victimizattion than thee NCVS. App
proach 2, how
wever, appeaars to track thhe NCVS rattes more
consisten
ntly across geographic and
d demograph
hic domains.
The estim
mates from both approach
hes are very sensitive
s
to sseries crimes and multiplee victimizatio
ons.
The core NCVS may be less sensitive to multip
ple victimizaations than thhe CS becausse of the 6-m
month
recall perriod: fewer reespondents reeport more than
t
one victtimization in a 6-month recall period tthan
in a 12-m
month period. In earlier yeears, series crrimes were exxcluded from
m victimizatio
on rates, but in
2011, thee number of incidents
i
rep
ported in a seeries crime (uup to 10), wass included in
n estimated
victimizattion rates. Fo
or example, for
f the Appro
oach 1 data, the area of C
Cook Countyy outside of
Chicago has
h an estimaated victimization rate off 62 if series ccrimes are exxcluded but tthe victimizattion
rate increeases to 98 if series crimess are counted
d as up to 100 incidents; thhe estimated violent
victimizattion rate for women is 833 without serries crimes annd 100 with sseries crimes. Tables 8-6 and
8-7 look at
a the percen
ntages of perssons in each domain whoo have had att least one viiolent
victimizattion.
NCVS-CS Pilot
P
Report
8-8
NCVS-CS Pilot Report
TTable 8-3.
Estimated
E
Violen
nt Victimization Rates,
R
per thoussand persons; 95
5 percent confideence intervals are given below the
estimates
e
A
All violent crime
S
Serious violent crime
TTable 8-4.
CS Approach 1
87
[57, 117]
51
[24, 82]
CS Ap
pproach 2
54
[35
5, 73]
20
[9
9, 30]
National NCVS
S, 2011
22.5
7.2
NCVS, Chicago
o CBSA,
2009-2011
15
[11, 19
9]
5
[3, 7]
Estimated
E
violent victimization ra
ates, per thousa
and persons, for geographic
g
posttstrata; 95 perce
ent confidence in
ntervals are
given
g
below the estimates
e
8-9
V
Violent crime, all areas
a
S
Serious violent crime, all areas
C
City of Chicago
Chicago, Low Crime
Chicago, Medium
m Crime
R
Remainder of Coo
ok County
R
Remainder of CBS
SA
CS Appro
oach 2
54
4
[35, 73]
7
20
0
[9, 30]
3
76
6
[38, 113]
1
47
7
[0, 94]
9
103
3
[26, 181]
1
86
6
[6, 16
65]
37
7
[9, 65]
6
50
0
[15, 85]
8
Chiicago
Police
e Dept.
1
10.6
3
3.8
7
7.7
21.8
Ratio of CS
Approach 1 to
NC
CVS, Chicago
CBSA
5.8
Ratio of CS
Approach 2 to
NCVS,
Ch
hicago CBSA
3.6
10.6
4.2
4.1
3.3
3.2
2.9
6.0
4.5
3.1
2.6
9.2
3.5
5.5
3.6
Estimation
Results
E
i
i R
l
Chicago, High Crrime
CS Approach 1
87
[57, 117]
51
[24, 82]
94
[58, 130]
52
[7, 97]
137
[61, 212]
106
[30, 181]
98
[16, 107]
76
[34, 118]
NCVS
S, Chicago
C
CBSA,
200
09-2011
15
[1
11, 19]
5
[3, 7]
23
[1
13, 33]
16
[[6, 26]
23
[[5, 40]
34
[[7, 60]
11
[[4, 17]
14
[[8, 19]
8
NCVS-CS Pilot Report
TTable 8-5.
Estimated
E
violent victimization ra
ates, per thousa
and persons, for demographic
d
doomains; 95 perce
ent confidence intervals
are
a given below the
t estimates
V
Violent crime, all
M
Men
W
Women
W
White nonHispanic
B
Black nonHispanicc
H
Hispanic
8-10
N
Never married
M
Married
A
Age 18-24
A
Age 25-34
A
Age 35-49
A
Age 65+
CS Approach 2
54
[35, 73]
73
[45, 101]
28
[15, 40]
56
[31, 82]
58
[10, 107]
42
[0, 85]
69
[32, 105]
27
[15, 40]
81
[10, 152]
67
[18, 115]
72
[19, 124]
33
[12, 55]
18
[5, 30]
National NCVS,
2011
1
22.5
5
Ra
atio of CS
App
proach 2 to
national NCVS,
2011
2.4
25.4
4
2
2.8
2.9
19.8
8
5
5.1
1.4
21.5
5
3
3.3
2.6
26.4
4
9
9.2
2.2
23.8
8
1
1.7
1.8
35.5
5
4
4.6
1.9
11.0
0
2
2.0
2.5
49.0
0
3
3.2
1.7
26.5
5
4
4.5
2.5
21.9
9
4
4.9
3.3
13.0
0
3
3.5
2.5
4.4
4
3
3.9
4.1
Estimation
Results
E
i
i R
l
A
Age 50-64
CS Approach 1
87
[57, 117]
71
[42, 101]
100
[49, 152]
71
[43, 100]
242
[63, 420]
40
[6, 73]
163
[88, 237]
22
[9, 35]
158
[57, 259]
119
[25, 212]
108
[38, 178]
45
[25, 65]
17
[3, 30]
Ratio
o of CS
Appro
oach 1 to
nation
nal NCVS,
20
011
3
3.9
8
NCVS-CS Pilot Report
TTable 8-6.
Estimated
E
rates of
o persons with at least one violent victimization
n, per thousand persons, for geo
ographic poststra
ata; 95
percent
p
confiden
nce intervals are given below the
e estimates
V
Violent crime, all areas
a
S
Serious violent crime, all areas
C
City of Chicago
Chicago, Low Crime
Chicago, Medium
m Crime
Chicago, High Crrime
R
Remainder of Coo
ok County
8-11
R
Remainder of CBS
SA
C Approach 1
CS
68
[48, 89]
39
[22, 57]
85
[51, 119]
50
[6, 94]
116
[56, 176]
99
[26, 172]
59
[14, 105]
64
[33, 94]
CS Approach 2
45
[31, 60]
19
[9, 29]
65
[32, 99]
24
[4, 44]
102
[24, 179]
83
[3, 162]
35
[7, 61]
38
[17, 59]
Estimation
Results
E
i
i R
l
8
Estimattion Results
Table 8-7
7.
Estima
ated numberr of persons with at leastt one violent victimization
n, per thousa
and
person
ns, for demographic dom
mains; 95 perrcent confide
ence intervalss are given b
below
the esstimates
CS Approach 1
68
[48, 89]
69
[39, 98]
69
[40, 98]
62
[40, 84]
146
[[42, 251 ]
40
[6, 73]
139
[78, 200]
20
[7, 33]
158
[57, 259]
91
[23, 158]
66
[40, 93]
42
[24, 61]
15
[3, 27]
Violent crime, all
Men
Women
White non
nHispanic
Black nonHispanic
Hispanic
Never marrried
Married
Age 18-24
4
Age 25-34
4
Age 35-49
9
Age 50-64
4
Age 65+
8.2
8
C
CS Approach 2
45
[31, 60]
66
[40, 93]
26
[14, 38]
44
[28, 61]
58
[10, 107]
37
[0, 78]
67
[31, 104]
27
[15, 40]
81
[10, 152]
63
[16, 110]
43
[20, 65]
33
[11, 54]
18
[5, 30]
Effectiiveness of
o Mail Sccreener ffor Prediccting NCV
VS
Victimization
Approach
h 2 involved sending selected househo
olds a short m
mail questionnnaire askingg about the
householld’s experiencces with crim
me in the prevvious 12 monnths. Questioons 9-12 on the mail
questionn
naire ask abo
out property crime,
c
and quuestions 13-116 ask about violent crim
me. Table 8-8 gives
the contin
ngency tables cross-classiifying househ
holds that repported at least one crime on the Apprroach
2 mail scrreener by wh
hether they haad at least on
ne in-scope N
NCVS crime in the teleph
hone intervieew.
NCVS-CS Pilot
P
Report
8-12
Estimattion Results
Table 8-8
8.
8
Cross--classification of reported
d crimes for h
households o
on mail scree
ener and NCVS
teleph
hone screene
er interview
Prope
erty Crime on Mail
M Screener
No to all of
o questions 9--12
Item nonre
esponse on so
ome of questions
9-12, no “yyes” answers
Yes to at least one of qu
uestions 9-12
Total
Violent Crime on Mail
M Screener
No to all of
o questions 13
3-16
Item nonre
esponse on so
ome of questions
13-16, no “yes” answerss
Yes to at least one of qu
uestions 13-16
Total
Anyy Crime on Ma
ail Screener
No to all of
o questions 9--16
Item nonre
esponse on so
ome of questions
9-16, no “yyes” answers
Yes to at least one of qu
uestions 9-16
Total
No
846
8
Properrty Crime on TTelephone Inte
erview
Yees
Total
108
8
954
0
8
146
1,000
164
4
272
2
310
1,272
Violen
nt Crime on Teelephone Interrview
No
Yees
Total
1,134
1
2
27
1,161
11
1
12
76
1,221
1
2
23
5
51
99
1,272
No
802
11
Any Crime on Teleephone Interview
Yees
11
15
1
157
970
18
86
30
02
Total
917
12
343
1,272
The Apprroach 2 mail screener hass relatively hiigh sensitivityy and specifiicity for deteccting crimes that
are subsequently repo
orted in the NCVS-2
N
instrrument, as shhown in Tablle 8-9.
Table 8-9
9.
Sensittivity and spe
ecificity of the Approach 2 Mail Scree
ener for NCVS
S crimes
Property Crime
C
Violent Criime
Any Crime
e
Sensitivity
S
0.60
0.45
0.62
Spe
ecificity
0
0.85
0
0.94
0
0.84
NOTE: In this table, the partial responses to the screenerr are counted a
as “No.”
The valuees in Table 8-9 for sensitiivity and speccificity are hiigh for victim
mization survveys, especiallly
when onee considers th
hat these queestions were not developeed for prediccting victimizzation but fo
or use
in establisshing subsam
mpling rates in
i a two-phasse approach. We fit logisttic regression
n models to 22005
NCVS puublic-use dataa, predicting whether a ho
ousehold hadd at least onee victimizatio
on from otheer
variables on the file suuch as demoggraphic variaables, incomee, professionn, and mobilitty. For thosee
models, the
t sensitivityy was only 0.28 for violen
nt crime and 0.22 for propperty crime w
when the
specificityy was set at 0.89.
0
NCVS-CS Pilot
P
Report
8-13
8
Estimattion Results
We further investigateed the incidents that weree misclassifieed by the Appproach 2 maail screener. T
Table
8-10 gives potential reeasons why households
h
answering “yees” to at least one of the crime screen
ner
questionss may have had
h no NCVSS-2 victimizattion reports. In Table 8-111, only the ffirst categoryy,
where no
o incidents reeported for th
he household
d on the NCV
VS-2 were inn scope for th
he survey,
provides clear evidencce of the missclassification
n; the adults aand childrenn who were n
not sampled ffor
the teleph
hone intervieew may have had victimizzations but thhere is no waay of knowin
ng for purposses of
this analyysis whether they
t
did or not.
n
Table 8-1
10.
Possib
ble reasons that
t
househo
olds reportingg crimes on tthe Approach
h 2 mail scre
eener
did no
ot have NCVS
S-2 victimizattion reports
Incidents reported, but none in scope
e for NCVS
HH subsam
mpled adults or
o at least one
e adult was no
onrespondentt
HH contains children wh
ho could have
e had victimiza
ations
At least on
ne of above
Total housseholds with crime
c
reported
d on mail screener but no in
nscope victimizations on telephone intterviews
All crrime
41
98
55
124
Prop
perty
crim
me
37
87
51
111
157
146
Violent
crime
14
46
31
62
76
We also looked
l
at thee misclassificaations where the mail scr eener reported no victim
mizations but the
subsequent telephonee interview haad at least on
ne victimizatiion report. O
One potentiall cause for suuch a
misclassiffication woulld be that thee person who
o filled out thhe mail screeener might haave been unaaware
of victimiizations occuurring to otheer household
d members. U
Unfortunatelly, we do nott know which
h
householld member filled out the mail
m screenerr so we cannot evaluate thhis hypothessis directly;
instead, th
he first row of
o Table 8-111 tallies the number
n
of hoouseholds wiith this type of
misclassiffication whicch have multiiple adults. We
W also do noot know the exact date off the incidentt; the
count in the
t second ro
ow of Table 8-11 is based
d on placing incidents at the 15th of tthe month.
Table 8-1
11.
House
eholds reportting victimiza
ation in the teelephone intterview but n
not in the ma
ail
screen
ner
All crrime
The house
ehold has multiple adults, so
o it is possible
e that the persson
returning the
t mail scree
ener is unaware of other vicctimizations
All inciden
nts reported occcurred after date
d
of mail screener
s
Incident re
eported on NC
CVS-2 was sim
mple assault orr threat
Incident re
eported on NC
CVS-2 was min
nor theft
At least on
ne of above
Total housseholds with at
a least one vicctimization reported in the
NCVS telephone intervie
ew but none re
eported in the
e Approach 2 m
mail
screener
NCVS-CS Pilot
P
Report
8-14
Prop
perty
crim
me
Violent
crime
8
84
2
28
6
8
9
98
7
76
2
25
22
6
6
8
8
89
24
11
16
10
08
28
Estimattion Results
8.3
8
Nonressponse Bias
B Asse
essment
In Section
n 8.1 we saw
w that the victtimization raates from the CS were higgher than tho
ose of the nattional
NCVS fo
or every type of crime. In this section we explore ppossible causses of the diff
fference. We
restrict th
he investigation to nonressponse bias in
n the CS; it iis possible thhat the core N
NCVS also has
bias but we
w do not exxplore that heere.
Tables 8--12 and 8-13 compare thee demograph
hic distributioon of the sam
mple to the diistribution off the
Chicago CBSA
C
using the 2009-20111 American
n Communityy Survey (AC
CS). Two verssions of weigghts
are used with
w the sam
mple. The base weights aree the inverse of the selecttion probabillities, so the
columns with the basee weights givve the demoggraphic distriibution of thee respondentts to the survvey.
The final weights are raked to geo
ographic area,, tenure, andd age/race/seex compositio
on as describ
bed in
Chapter 7,
7 so the coluumns with th
he final weigh
hts show how
w well the noonresponse w
weighting
performs on characterristics known
n from the ACS.
A
Using thhe base weighhts, it is seen
n that the sam
mple
from eith
her approach has substanttial underrepresentation oof renters, off young adultts, of men, an
nd of
minorities. These charracteristics were
w used as poststratifica
p
ation and rakking factors in
n constructin
ng the
final weigghts, so the estimates
e
of these
t
characteristics usingg the final weeights corresp
pond exactlyy to
the Amerrican Commuunity Survey estimates.
The base-weighted sample has an overrepresen
ntation of higgh income households, m
married or
widowed persons, and
d highly educcated persons. The base-w
weighted sam
mple also is leess mobile th
han
the CBSA
A population
n as a whole: Table
T
8-13 compares thee percentagess of adults in the ACS wh
ho did
not live in
n the house or
o apartmentt one year ago
o with the peercentages off respondentts to the CS w
who
lived in th
he dwelling unit
u less than
n one year. Th
he final weigghts bring thee estimates of marital stattus,
householld income, ed
ducation, and
d mobility clo
oser to the A
ACS values, bbut do not co
ompletely corrrect
for the diiscrepancies. Note that th
he unweighteed national coore NCVS allso has underrrepresentatio
on of
these groups compareed to nationaal totals, but to
t a lesser deegree than thhe CS.
The undeerrepresented
d groups in th
he CS and NCVS
N
are thoose that tend to have the h
highest
victimizattion rates. Evven using thee base weights, however, the CS exhibbits higher viictimization rrates
than the NCVS.
N
The base-weighte
b
ed property victimization
v
rates for thee CS are 261 per thousandd
househollds for Appro
oach 1 and 275 per thoussand househoolds for Approach 2. Thee base-weightted
violent viictimization rates
r
are 58 per
p thousand
d persons forr Approach 1 and 41 per tthousand perrsons
for Appro
oach 2. The property
p
victtimization rates calculatedd using the bbase weights are in fact veery
close to those calculatted using thee final nonressponse-adjus ted weights ((273 for App
proach 1 and 287
NCVS-CS Pilot
P
Report
8-15
8
Estimattion Results
for Appro
oach 2), and still higher th
han the NCV
VS estimate ffor the Chicaago CBSA off 119. The vio
olent
victimizattion rates aree somewhat more
m
sensitivve to the nonnresponse addjustments in the final weiights:
the violen
nt victimizatiion rates calcculated using final weightss are 87 for A
Approach 1 aand 54 for
Approach
h 2 (compareed to the NC
CVS estimate of 15). This is consistentt with the ressults in Sectio
on
8.1 showiing that the estimated
e
vicctimization raates from thee CS were higgher than thee NCVS for every
subgroup
p examined.
Table 8-1
12.
Comparison of NCVS-CS Pilot household
h
in terview resp
pondents with
h population
Percent off
occupied
householdss
from ACS
Geograph
hic Area
City of Ch
hicago
Low
Medium
m
High
Remain
nder Cook
Remain
nder CBSA
Total
Tenure
Owned
Rented
Total
Household Income
< $25,0
000
$25,00
00 - $49,999
$50,00
00-$74,999
$75,00
00+
NCVS-CS Pilot
P
Report
Approach 1
Telephon
ne Screener
Base
weights
Final
weights
Approa
ach 2
Mail Screeener
Base
weights
Final
weights
Approach 2
TTelephone Scrreener
Base
w
weights
Final
w
weights
12.2%
8.5%
9.2%
26.7%
43.4%
100.0%
11.7%
7.7%
6.8%
27.7%
46.1%
100.0%
12.2%
8.5%
9.2%
26.7%
43.4%
100.0%
12.1%
7.4%
7.0%
28.1%
45.5%
100.0%
12.2%
8.5%
9.2%
26.7%
43.4%
100.0%
12.0%
7.1%
8.2%
28.2%
44.5%
1
100.0%
1
12.2%
8.5%
9.2%
2
26.7%
4
43.4%
10
00.0%
66.0%
34.0%
100.0%
75.8%
24.2%
100.0%
66.0%
34.0%
100.0%
77.3%
22.7%
100.0%
66.0%
34.0%
100.0%
77.5%
22.5%
1
100.0%
6
66.0%
3
34.0%
10
00.0%
20.7%
21.9%
18.1%
39.3%
100.0%
7.5%
23.6%
22.0%
46.9%
100.0%
9.9%
26.1%
21.8%
42.2%
100.0%
6.9%
23.3%
23.0%
46.8%
1
100.0%
9.0%
2
26.1%
2
22.9%
4
42.1%
10
00.0%
8-16
Estimattion Results
Table 8-1
13.
8
Comparison of NCVS-CS Pilot person
p
intervview respond
dents with po
opulation
Percent of
Ag
ge 18+
Pop
pulation
from
m ACS
Age
4
18-24
25-34
4
35-44
4
45-54
4
55-64
4
65+
Total
Approacch 1
Base
Weights
W
Final
Weights
Ap
pproach 2
Base
Weights
Fina
al
Weigh
hts
12
2.6%
19
9.3%
18
8.6%
19
9.5%
14
4.8%
15
5.2%
100
0.0%
4.0%
11.9%
13.7%
18.8%
20.1%
31.6%
100.0%
12.6%
19.3%
18.6%
19.5%
14.8%
15.2%
100.0%
3.3%
%
12.3%
%
14.7%
%
19.5%
%
22.5%
%
27.9%
%
100.0%
%
12..6%
19..3%
18..6%
19..5%
14..8%
15..2%
100..0%
51
1.9%
48
8.1%
100
0.0%
55.4%
44.6%
100.0%
51.9%
48.1%
100.0%
54.7%
%
45.3%
%
100.0%
%
51..9%
48..1%
100..0%
10.7%
12.3%
77.0%
100.0%
16.4%
17.8%
65.8%
100.0%
11.0%
%
12.0%
%
77.0%
%
100.0%
%
16..4%
17..8%
65..8%
100..0%
Marital Status
Marrie
ed
Widow
wed
Divorcced/Separated
d
Neverr Married
Total
16
6.4%
17
7.8%
65
5.8%
100
0.0%
(age 15+)
47.6%
4
5.7%
11.0%
1
35.7%
3
10
00.0%
57.1%
10.4%
11.7%
20.8%
100.0%
51.3%
6.3%
10.6%
31.8%
100.0%
56.7%
%
10.9%
%
11.9%
%
20.5%
%
100.0%
%
48.3%
6.8%
11
1.4%
33.6%
100.0%
Education
nal Attainmen
nt
10%).
(
Thiss question iis long. Could be broke
en into 2
questions,
q
ending
e
the first
f
after “b
broken into yyour home.”
b.
illegally gotte
en in or tried to get into
o a garage, shed, or sto
orage room?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
c.
illegally gotte
en in or tried to get into
o a hotel or motel room
m or vacatio
on home wh
here
you were
e staying?
hing notewo
orthy based
d on selecte d interviewss.
Notes: Noth
NCVS-CS
S Pilot CATI Screen Nam
me: B37B
[BREAK IN.] How ma
any times did
d incidentss of this typ e happen to
o you in the last 12
months?
?
hing notewo
orthy based
d on selecte d interviewss.
Notes: Noth
NCVS-CS Pilot
P
Report
D-3
Appen
ndix D
Behavior Codingg Questions Reeviewed With Recommenda
ations
NCVS-CS
S Pilot CATI Screen Nam
me: B37C
[BREAK--IN IN THE LA
AST 12 MONTHS]
What ha
appened? Pllease brieflyy describe the [most reecent/next] incident.
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
NCVS-CS
S Pilot CATI Screen Nam
me: B38
What wa
as the total number of cars,
c
vans, trucks,
t
mottorcycles, orr other moto
or vehicles
owned by
b you or anyy other mem
mber of thiss household
d during the last 12 mo
onths? Include
those yo
ou no longerr own.
Notes: Mosst problema
atic question
n in NCVS-1
1, from a beh
havioral cod
ding
perspective
e. 16 majorr misreads, usually leavving out “an
ny other member of thiss
household”” or “Include
e those you no longer oown.” R inte
errupted or a
asked for
clarification 24 times, interviewerr probed or offered clarification 20
0 times. Thiis is
mplicated qu
uestion with multiple qu
ualificationss. Passive voice and “w
what
a very com
was the tottal number”” rather than
n “how man
ny” add to th
he complexxity.
NCVS-CS
S Pilot CATI Screen Nam
me: B39A
During th
he last 12 months,
m
(oth
her than the
e incidents a
already men
ntioned,) ...
a.
(w
was the veh
hicle/were any
a of the ve
ehicles) stollen or used without perrmission?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
b.
did anyone steal
s
any pa
arts such as a tire, car sstereo, hubccap, or batte
ery?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
c.
did anyone steal
s
any gas from (it/th
hem)?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
d.
did anyone attempt
a
to steal
s
any vehicle or parrts attached
d to (it/them
m)
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
NCVS-CS Pilot
P
Report
D-4
Appen
ndix D
Behavior Codingg Questions Reeviewed With Recommenda
ations
S Pilot CATI Screen Nam
me: B37B
NCVS-CS
[VEHICLE
E INCIDENT..] How many times did incidents oof this type h
happen to yyou in the la
ast
12 montths?
hing notewo
orthy based
d on selecte d interviewss.
Notes: Noth
NCVS-CS
S Pilot CATI Screen Nam
me: B39C
[VEHICLE
E INCIDENT IN THE LAS
ST 12 MONTTHS] What h
happened? Please briefly describe
e the
[most re
ecent/next] incident.
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
NCVS-CS
S Pilot CATI Screen Nam
me: B40A1
(Other th
han the incid
dents already mentioned,) Since ( DATE ONE Y
YEAR AGO),, were you
attacked
d or threatened or did you
y have something stoolen from yo
ou...
a.
at home including the porch
p
or yard
d?
hing notewo
orthy based
d on selecte d interviewss.
Notes: Noth
b.
at or near a friend’s,
f
rela
ative’s, or neighbor’s home?
Notes:
N
Re-re
eading requ
uired 8 times out of 83.. Syntax awkward; “at o
or near
someone
s
else’s home, such as tha
at of a friend
d, relative, o
or neighborr?”
c.
at work or scchool?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
d.
in
n places succh as a stora
age shed orr laundry rooom, a shopp
ping mall, restaurant,
bank, or airport?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
NCVS-CS Pilot
P
Report
D-5
Appen
ndix D
Behavior Codingg Questions Reeviewed With Recommenda
ations
NCVS-CS
S Pilot CATI Screen Nam
me: B40A2
(Other th
han the incid
dents already mentioned,) Since { DATE ONE Y
YEAR AGO},, were you
attacked
d or threatened or did you
y have something stoolen from yo
ou...
e.
while
w
riding in
i any vehiccle?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
f.
on the streett or in a parking lot?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
g.
at such place
es as a partty, theater, gym,
g
picnic area, bowliing lanes, or while fishiing
or huntin
ng?
Notes: 8 major
m
misrea
ads – mostlyy omission of one or m
more examples. Can be
awkward to
o read to older Respon
ndents.
NCVS-CS
S Pilot CATI Screen Nam
me: B40A3
We have
e just discusssed times when
w
someone may ha
ave attacked, threatene
ed, or stolen
n
something from you
u. We’re also interested
d in times w
when someo
one may have tried to
attack, threaten, or steal some
ething from you.
(Other th
han the incid
dents already mentioned,) Since ( DATE ONE Y
YEAR AGO),, did anyone
e
attempt to attack or attempt to
o steal anything belongging to you ffrom any off these places?
Notes: 36%
% mis-readss, a variety of
o minor thin
ngs.
NCVS-CS
S Pilot CATI Screen Nam
me: B40B
[ATTACK
K OR THEFT OR ATTEMP
PT OF ATTAC
CK OR THEFFT.] How ma
any times did incidentss of
this type
e happen to you in the last 12 mon
nths?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
NCVS-CS Pilot
P
Report
D-6
Appen
ndix D
Behavior Codingg Questions Reeviewed With Recommenda
ations
NCVS-CS
S Pilot CATI Screen Nam
me: B40C
[ATTACK
K OR THEFT OR ATTEMP
PT OF ATTAC
CK OR THEFFT IN THE LA
AST 12 MON
NTHS] Whatt
happene
ed? Please briefly
b
describe the [mo
ost recent/n
next] incide
ent.
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
S Pilot CATI Screen Nam
me: B41A1
NCVS-CS
During th
he last 12 months,
m
(oth
her than the
e incidents a
already men
ntioned,) ha
as anyone
attacked
d or threatened you in any
a of these
e ways...
[EXCLUD
DE TELEPHO
ONE THREATTS.]
a.
with
w any wea
apon, for insstance, a gu
un or knife??
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
b.
with
w anything
g like a basseball bat, a frying pan,, scissors, or stick?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
c.
by something
g thrown, su
uch as a rocck or bottle??
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
d.
in
nclude any grabbing,
g
pu
unching, or choking?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
NCVS-CS Pilot
P
Report
D-7
Appen
ndix D
Behavior Codingg Questions Reeviewed With Recommenda
ations
NCVS-CS
S Pilot CATI Screen Nam
me: B41A2
During th
he last 12 months,
m
(oth
her than the
e incidents a
already men
ntioned,) ha
as anyone
attacked
d or threatened you in any
a of these
e ways...
[EXCLUD
DE TELEPHO
ONE THREATTS.]
e.
any rape, atttempted rap
pe or other type
t
of sexu
ual attack?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
f.
any face to face threatss?
hing notewo
orthy based
d on selecte d interviewss.
Notes: Noth
g.
any attack or threat or use
u of force by anyone at all? Plea
ase mention
n it even if yyou
c
it wa
as a crime.
are not certain
Notes: 36%
% misreads, a majority major – om
mission of la
ast sentence
e most
frequent.
NCVS-CS
S Pilot CATI Screen Nam
me: B41B
[ATTACK
K, THREAT, OR
O USE OF FORCE.]
F
How
w many tim
mes did incid
dents of thiss type happ
pen
to you in
n the last 12
2 months?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
NCVS-CS
S Pilot CATI Screen Nam
me: B41C
[ATTACK
K, THREAT, OR
O USE OF FORCE
F
IN TH
HE LAST 12
2 MONTHS] What happ
pened? Plea
ase
briefly de
escribe the [most recen
nt/next] inccident.
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
NCVS-CS Pilot
P
Report
D-8
Appen
ndix D
Behavior Codingg Questions Reeviewed With Recommenda
ations
NCVS-CS
S Pilot CATI Screen Nam
me: B42A
People often
o
don’t think
t
of incidents comm
mitted by soomeone the
ey know. Du
uring the lasst 12
months, (other than
n the inciden
nts already mentioned,,) did you ha
ave someth
hing stolen ffrom
or were you
y attacke
ed or threate
ened by...
[EXCLUD
DE TELEPHO
ONE THREATTS.]
a.
so
omeone at work or sch
hool?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
b.
a neighbor or friend?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
c.
a relative or family mem
mber?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
d.
any other person you’ve
e met or kno
own?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
NCVS-CS
S Pilot CATI Screen Nam
me: B42B
[ATTACK
K OR THREA
AT BY SOMEONE YOU KNOW.] How
w many time
es did incide
ents of this ttype
happen to
t you in the
e last 12 months?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
NCVS-CS
S Pilot CATI Screen Nam
me: B42C
[ATTACK
K OR THREA
AT BY SOMEONE YOU KNOW IN THE
E LAST 12 M
MONTHS] W
What
happene
ed? Please briefly
b
describe the [mo
ost recent/n
next] incide
ent.
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
NCVS-CS Pilot
P
Report
D-9
Appen
ndix D
Behavior Codingg Questions Reeviewed With Recommenda
ations
NCVS-CS
S Pilot CATI Screen Nam
me: B43A
Incidentss involving forced
f
or un
nwanted sexxual acts arre often difficult to talk about. During
the last 12
1 months,, {(other than the incide
ents alreadyy mentioned
d,) have you
u been force
ed or
coerced to engage in
i unwanted
d sexual acttivity by...
a.
so
omeone you
u didn’t kno
ow?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
b.
a casual acquaintance?
?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
c.
so
omeone you
u know welll?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
NCVS-CS
S Pilot CATI Screen Nam
me: B43B
[FORCED
D OR UNWA
ANTED SEXU
UAL ACTS.] How many times did in
ncidents of tthis type
happen to
t you in the
e last 12 months?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
NCVS-CS
S Pilot CATI Screen Nam
me: B43C
[FORCED
D OR UNWA
ANTED SEXU
UAL ACTS IN
N THE LAST 1
12 MONTHS
S] What ha
appened?
Please briefly
b
descrribe the [mo
ost recent/n
next] inciden
nt.
Notes: Too much prob
bing by one interviewer
i
at this quesstion (rathe
er than waitiing
to collect th
he details la
ater in the NCVS2
N
instrrument).
NCVS-CS Pilot
P
Report
D-10
Appen
ndix D
Behavior Codingg Questions Reeviewed With Recommenda
ations
NCVS-CS
S Pilot CATI Screen Nam
me: B44A
During th
he last 12 months,
m
(oth
her than the
e incidents a
already men
ntioned,) did you call th
he
police to
o report som
mething thatt happened to you whicch you thought was a ccrime?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
S Pilot CATI Screen Nam
me: B44C
NCVS-CS
During th
he last 12 months,
m
werre you attaccked or threeatened, or w
was someth
hing stolen or
an attem
mpt made to
o steal something that belonged too you or ano
other house
ehold memb
ber?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
NCVS-CS
S Pilot CATI Screen Nam
me: B44D
[CALL PO
OLICE TO RE
EPORT CRIM
ME AND ATTTACK, THRE
EAT, THEFT, OR THEFT A
ATTEMPT ON
N
RESPON
NDENT OR HOUSEHOLD
H
D MEMBER.] How manyy times did incidents off this type
happen to
t you in the
e last 12 months?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
NCVS-CS
S Pilot CATI Screen Nam
me: B44C
[CALL PO
OLICE TO RE
EPORT CRIM
ME AND ATTTACK, THRE
EAT, THEFT, OR THEFT A
ATTEMPT ON
N
RESPON
NDENT OR HOUSEHOLD
H
D MEMBER IN THE LASTT 12 MONTH
HS] What h
happened?
Please briefly
b
descrribe the [mo
ost recent/n
next] inciden
nt.
Notes: Nothing note
eworthy base
ed on seleccted intervie
ews.
NCVS-CS
S Pilot CATI Screen Nam
me: B45A
During th
he last 12 months,
m
(oth
her than the
e incidents a
already men
ntioned,) did anything
which yo
ou thought was
w a crime
e happen to you, but yoou did not re
eport to the police?
Notes: Noth
hing notewo
orthy based
d on selecte d interviewss.
NCVS-CS Pilot
P
Report
D-11
Appen
ndix D
Behavior Codingg Questions Reeviewed With Recommenda
ations
NCVS-CS
S Pilot CATI Screen Nam
me: B45C
During th
he last 12 months,
m
werre you attaccked or threeatened, or w
was someth
hing stolen or
an attem
mpt made to
o steal something that belonged too you or ano
other house
ehold memb
ber?
Notes: Nothing note
eworthy base
ed on seleccted intervie
ews.
S Pilot CATI Screen Nam
me: B45D
NCVS-CS
[CALL PO
OLICE TO RE
EPORT CRIM
ME AND ATTTACK, THRE
EAT, THEFT, OR THEFT A
ATTEMPT ON
N
RESPON
NDENT OR HOUSEHOLD
H
D MEMBER.] How manyy times did incidents off this type
happen to
t you in the
e last 12 months?
Notes: Nothing note
eworthy base
ed on seleccted intervie
ews.
NCVS-CS
S Pilot CATI Screen Nam
me: B45C
[DID NOTT CALL POLICE TO REP
PORT CRIME
E IN THE LAS
ST 12 MONTTHS] What happened??
Please briefly
b
descrribe the [mo
ost recent/n
next] inciden
nt.
Notes: Nothing note
eworthy base
ed on seleccted intervie
ews.
NCVS-CS Pilot
P
Report
D-12
Appendix E
Census Flow Chart of TOC Codes
Appendix E
Census Flow Chart of TOC C
Codes
NCVS-CS Pilot
P
Report
E-1
Appendix E
Census Flow Chart of TOC C
Codes
NCVS-CS Pilot
P
Report
E-2
Appendix E
Census Flow Chart of TOC C
Codes
NCVS-CS Pilot
P
Report
E-3
Appendix E
Census Flow Chart of TOC C
Codes
NCVS-CS Pilot
P
Report
E-4
Appendix E
Census Flow Chart of TOC C
Codes
NCVS-CS Pilot
P
Report
E-5
Appendix E
Census Flow Chart of TOC C
Codes
NCVS-CS Pilot
P
Report
E-6
Appendix E
Census Flow Chart of TOC C
Codes
NCVS-CS Pilot
P
Report
E-7
Appendix E
Census Flow Chart of TOC C
Codes
NCVS-CS Pilot
P
Report
E-8
Appendix E
Census Flow Chart of TOC C
Codes
NCVS-CS Pilot
P
Report
E-9
Appendix E
Census Flow Chart of TOC C
Codes
NCVS-CS Pilot
P
Report
E-10
Appendix E
Census Flow Chart of TOC C
Codes
NCVS-CS Pilot
P
Report
E-11
Appendix F
Type of Crime Specifications
Appe
endix F
Type of C
Crime Specifica
ations
The tablee below provvides the speccifications for Census Typpe of Crime (TOC) Codiing. Codes aare
presented
d in programming order (and
(
are not necessarily
n
liisted in numeeric order).
Situatio
on
TOC=00 and
a SC634a=1
1 and
(SC646_0
01=1 or SC655
5_02 = 1)
TOC=00 and
a SC634a = 1 and
(SC655_0
03=1 or SC646
6_02=1 or
(SC643_0
01=1 and
any SC643
3_07-13=1)))
TOC reccode
01
1
02
TO
OC
descrription
Compl eted
Rape
Attemp
pted
Rape
TOC=00 and
a SC634a=1
1 and
(SC646_0
03=1 or SC655
5_04=1 or
SC643_05
5=1 or SC643
3_06=1) and (a
any
SC638_01
1-06=1 or SC6
638LIC=3 or 7 or
any SC655
5_04-09=1 orr any SC643_0
0711=1 or (((SC655_10=1
1 or SC655_11
1=1
or SC655=
=yes ‘NA’) and
d SC663=2-20
00))
03
Sexuall Attack
with Seerious
Assaullt
TOC=00 and
a SC634a=1
1 and
(SC646_0
03=1 or SC655
5_04 = 1 or an
ny
SC643_05
5-06=1) and (any SC655_1011=1 or SC655=yes
S
‘NA
A’) or (any
SC643_05
5-06=1 and
any SC643
3_12-13=1))
04
Sexuall Attack
with M
Minor
Assaullt
TOC=00 and
a SC634a=1
1 and SC731a
a=1
and
(any SC65
55_5-9=1 or
((any SC65
55_10-11 or SC655=yes
S
‘N
NA’)
and ((any SC638_1-6 or SC638LIC=3
3 or
7) or ((SC6
655_11=1 or SC655=yes ‘N
NA’)
and SC663 = 2-200))))
05
Compl eted
Robbeery with
Injury ffrom
Seriou s
Assaullt
TOC=00 and
a SC634a=1
1 and SC732a
a=1
and
(any SC65
55_5-9=1 or (((any SC655_1
1011=1 or SC655=yes
S
‘NA
A’) and ((any
SC638_1--6=1 or SC638
8LIC=3,7) or
((SC655_1
11=1 or SC65
55=yes ‘NA’) and
SC663=2--200))))
08
Attemp
pted
Robbeery with
Injury ffrom
Seriou s
Assaullt
NCVS-CS Pilot
P
Report
F-1
Recode d
description
Presence and rap
ped (howattacck,
inju
ury)
Presence and (atttempted rape
e
(how
wattack, injurry) or (verbal th
hreat
of rape with (wea
apon or being
g
follo
owed, or strucck w/o weapo
on
(how
wtryattack/th
hreaten))))
exual assault o
other
Presence and (se
empted rape
than rape or atte
(how
wattack, injurry), or unwante
ed
sexual contact w//or w/o force
(how
wtryattack/th
hreaten)) with (any
typee of weapon o
or weapon pre
esent
(weeapon) or seve
ere injury not rrape
or a
attemped rape
e (injury), or le
esser
inju
ury with hospittal stay of 2 orr
more days (injuryy and
careedayhospital)))
exual assault, not
Presence and (se
ed rape(howatttack,
rapee or attempte
inju
ury) or unwantted sexual con
ntact
with
h or w/o force
e
(how
wtryattack/th
hreaten)) and
(minor injury (inju
ury) or yes NA
A in
inju
ury, or (unwantted sexual con
ntact
w/ or w/o force a
and (followed,, hit,
slap
pped, surrounded, etc.
(how
wtryattack/th
hreaten))))
Presence, Theft, and
onsexual majo
or injury(injury)) or
(No
A’ in
((minor injury (injjury) or yes ‘NA
ury) and
inju
(Anyy weapon present (weapon) or
((other specified injury or yes ‘NA’
(inju
ury)) and hosp
pital stay of 2--200
dayys))))
Presence, Attempted theft, and
nsexual major injury(injury)) or
(non
((minor injury (injjury) or yes ‘NA
A’ in
inju
ury) and (any w
weapon present
(weeapon) or ((oth
her specified in
njury
or yyes ‘NA’ (injuryy)) and hospita
al
stayy of 2-200 dayys
(carredayhospit))))))
Appe
endix F
Type of C
Crime Specifica
ations
TO
OC
descrription
Aggravvated
Assaullt
Compl eted
njury
with In
Situatio
on
TOC=00 and
a SC634a = 1 and
(any SC65
55_5-9=1 or (((any SC655_1
1011=1 or SC655=yes
S
‘NA
A’) and ((any
SC638_1--6=1 or SC638
8LIC=3,7) or
((SC655_1
11=1 or SC65
55=yes ‘NA’) and
SC663=2--200))))
TOC reccode
1
11
TOC=00 and
a SC634a=1
1 and SC731a
a=1
and
(SC655_1
10=1 or SC655
5_11=1 or
SC655=ye
es ‘NA’)
TOC=00 and
a SC634a=1
1 and SC732a
a=
1 and (anyy SC655_10-1
11=1 or
SC655=ye
es ‘NA’)
06
TOC=00 and
a SC634a=1
1 and
(any SC65
55_10-11=1 or
o SC655=yes
‘NA’)
14
TOC=00 and
a SC634a=1
1 and SC731a
a=1
and (any SC646_04,06
S
,07,09,10,12-14=1 or SC646=yes
S
‘NA
A’ or any
SC643_02
2-03,07-14=1
1 or SC643=ye
es
‘NA’)
07
TOC=00 and
a SC634a=1
1 and SC732a
a=1
and
(any SC64
46_04,06,07,0
09,10,12-14=
=1
or SC646=
=yes ‘NA’ or any SC643_02
203,07-14=
=1 or SC643=
=yes ‘NA’)
10
Attemp
pted
Robbeery
withou
ut Injury
TOC=00 and
a SC634a=1
1 and
(any SC63
38_1-6=1 or any
a
SC638LIC
C=3,7) and (an
ny SC643_0813=1 or SC643=yes
S
‘NA
A’ or any
SC646_06
6,07,09,10,12
2-14=1 or
SC646=ye
es ‘NA’)
TOC=00 and
a SC634a=1
1 and
(any SC63
38_1-6=1 or SC638LIC=3,7
S
)
12
Attemp
pted
Aggravvated
Assaullt with
Weapoon
13
Threattened
Assaullt with
Weapoon
Sexuall
Assaullt
withou
ut Injury
TOC=00 and
a SC634a=1
1 and
(SC646_0
03=1 or SC643
3_05=1)
NCVS-CS Pilot
P
Report
Compl eted
Robbeery with
Injury ffrom
Minor A
Assault
Attemp
pted
Robbeery with
Injury ffrom
Minor A
Assault
Simplee
Assaullt
Compl eted
with In
njury
Compl eted
Robbeery
withou
ut Injury
09
15
F-2
Recode d
description
onsexual majo
or
Presence and (no
inju
ury(injury) or (((minor injury
(inju
ury) or yes ‘NA
A’ in injury) an
nd
(anyy weapon pressent (weapon)) or
((other specified injury or yes ‘NA’
(inju
ury)) and hosp
pital stay of 2--200
dayys (caredayhosspit)))))
Presence, Theft, and (minor injjury
(inju
ury) or yes ‘NA
A’ in injury)
Presence, Attempted theft, and
(minor injury (inju
ury) or yes ‘NA
A’ in
inju
ury)
minor injury (injjury)
Presence and (m
ury)
or yyes ‘NA’ in inju
Presence, Theft, and (shot, hit with
an o
object or weap
pon, stabbed, hit,
grabbed, or otherr specify
(how
wattack) or ye
es ‘NA’ in
how
wattack or yess ‘NA’ for
how
wattack/threa
aten or any typ
pe of
atteempted attack
k or threat
excluding ones off a sexual natture
(how
wtryattack/th
hreaten)
Presence, Attempted theft, and
(sho
ot, hit with an object or wea
apon,
stab
bbed, hit, grab
bbed, or otherr
speecify (howattacck) or yes ‘NA’ in
how
wattack or yess ‘NA’ for
how
wattack/threa
aten or any typ
pe of
atteempted attack
k or threat
excluding ones off a sexual natture
(how
wtryattack/th
hreaten)
eapon presentt
Presence and we
hot at, attemp
pted
(weeapon) and (sh
atta
ack , or attack
k with or w/o
wea
apon (howatta
ack,
how
wtryattack/thrreaten)))
eapon presentt
Presence and we
(weeapon)
Presence and unwanted sexua
al
assault or contacct with force–n
not
rapee or attempte
ed rape–
(how
wattack, howttryattack/thre
eaten)
Appe
endix F
Type of C
Crime Specifica
ations
Situatio
on
TOC=00 and
a SC634a=1
1 and
SC643_06
6=1
TOC reccode
16
TOC=00 and
a SC634a = 1 and
(SC646_1
10,12-14=1 orr SC646=yes ‘NA’
or (any SC
C643_02-03 = 1 and any
SC643_12
2-13=1) or SC
C643_11,1314=1 or SC643=yes
S
‘NA
A’)
17
TOC=00 and
a SC634a = 1 and
SC643_01
1=1
TOC=00 and
a SC634a=1
1 and SC643_
_04
=1
18
TO
OC
descrription
Unwan
nted
Sexuall
Contacct
withou
ut Force
Assaullt
withou
ut
Weapoon
withou
ut Injury
Verbal Threat
of Rap
pe
Verbal Threat
of Sexu
ual
Assaullt
Verbal Threat
of Assa
ault
19
TOC=00 SC634a
S
= 1 an
nd
any SC643
3_02-03=1
20
TOC=00 and
a SC634a =1 and SC767a
a=1
and SC748_02=1
21
1
TOC=00 and
a SC634a=1
1 and SC767a
a=1
and (SC74
47=1-99996 or
o any
SC748_01
1,03=1)
23
TOC=00 and
a SC634a=1
1 and
sc733_02
2=1 and sc742
2a=1
22
Attemp
pted
Purse
Snatch
hing
TOC=00 and
a SC625LIC=
=1 and SC618
8=
1
31
1
TOC=00 and
a SC625LIC=
=1
33
Compl eted
Burgla
ary,
Forciblle Entry
Attemp
pted
Forciblle Entry
TOC=00 and
a SC618=1
32
TOC=00 and
a SC732a = 1 and
(SC733_0
05=1 or SC733
3_06=1)
41
1
NCVS-CS Pilot
P
Report
Compl eted
Purse
Snatch
hing
Pockett
Pickingg
(Comp
pleted
only)
Compl eted
Burgla
ary,
Unlawfful
Entry W
Without
Force
Attemp
pted
Motor Vehicle
Theft
F-3
Recode d
description
Presence and unwanted sexua
al
con
ntact without fforce
(how
wtryattack/th
hreaten)
Presence and (hit by thrown ob
bject,
slap
pped, grabbed
d, other, etc. o
or
yes’NA’ (howattack) or (verbal
threeat of any type
e other than ra
ape
or ssexual attack, and followed,
surrrounded, tried
d to it, slap, etc.) or
objeect thrown at person, tried tto hit,
slap
p, other, etc., o
or yes “na”
(how
wtryattack/th
hreaten)
Presence and verrbal threat of rape
(how
wtryattack/th
hreaten)
Presence and verrbal threat of
e
sexual assault otther than rape
(how
wtryattack/th
hreaten)
her
Presence and verrbal threat oth
than to rape or se
exually assault
(how
wtryattack/th
hreaten)
Presence, purse sstolen
hatwastaken), and cash on
(wh
persson (cashonpe
erson)
on person
Presence, cash o
and (cash tak
ken
(casshonperson), a
betw
ween 1-99996
6
(am
mountcashtake
en) or cash or
wallet taken (wha
atwastaken))
purse
Presence, Attempted to steal p
at), the purse was
(atttempttheftwha
on tthe person
(atttempttheftonp
person)
Theere was eviden
nce that the
offeender got in byy force (eviden
nce,
offeenderinside)
Theere was eviden
nce that the
offeender got in orr tried to get in by
forcce (evidence)
the offender got inside
(offfenderinside)
Thee offender atte
empted to tak
ke car
or o
other motor ve
ehicle from
resp
pondent or hh
hmemember
(atttempttheft, atttempttheftwh
hat)
Appe
endix F
Type of C
Crime Specifica
ations
Situatio
on
TOC=00 and
a SC732a = 1 and any
SC616 =0
01-02,05-07
TOC reccode
39
TOC=00 and
a SC732a=1
1
29
TOC=00 and
a SC731a = 1 and any
SC748_05
5-06=1 and (a
any SC763=2,3
or SC764=
=2)
40
TOC=00 and
a SC731a=1
1 and (SC748=
=14,7-27 or SC748=yes ‘N
NA’) and SC61
16
= 1-2,5-7 and sum<10
34
TOC=00 and
a SC731a=1
1 and (SC748=
=14,7-27 or SC748=yes ‘N
NA’) and SC61
16
= 1-2,5-7 and sum<50
35
TOC=00 and
a SC731a=1
1 and (SC748=
=14,7-27 or SC748=yes ‘N
NA’) and SC61
16
= 1-2,5-7 and sum<250
0
36
TOC=00 and
a SC731a=1
1 and (SC748=
=14,7-27 or SC748=yes ‘N
NA’) and SC61
16
= 1-2,5-7 and sum=999
998
TOC=00 and
a SC731a=1
1 and (SC748=
=14,7-27 or SC748=yes ‘N
NA’) and SC61
16
= 1-2,5-7
TOC=00 and
a SC731a=1
1 and (SC748=
=14,7-27 or SC748=yes ‘N
NA’) and sum<
<10
TOC=00 and
a SC731a=1
1 and (SC748=
=14,7-27 or SC748=yes ‘N
NA’) and sum<
<50
NCVS-CS Pilot
P
Report
TO
OC
descrription
Attemp
pted
House hold
Larcen
ny
Attemp
pted
Person
nal
Larcen
ny
withou
ut
Contacct
Compl eted
Motor Vehicle
Theft
Compl eted
House hold
Larcen
ny (less
than $
$10)
Compl eted
House hold
Larcen
ny ($10$49)
Compl eted
House hold
Larcen
ny ($50$249)
Compl eted
House hold
Larcen
ny
(Value NA)
Compl eted
House hold
Larcen
ny
($250+
+)
38
37
Compl eted
Person
nal
Larcen
ny
withou
ut
Contacct (less
than $
$10)
Compl eted
Person
nal
Larcen
ny
withou
ut
Contacct ($10$49)
24
25
F-4
Recode d
description
Atteempted theft in or near the
hom
me– detached
d building on
property, yard, ap
partment hall,, or on
o own home
streeet adjacent to
(atttempttheft, loccation_1)
Atteempted theft
(atttempttheft)
Theeft of car or mo
otor vehicle w
where
no//don’t know if permission w
was
giveen to offenderr to use vehicle, or
offeender did not return car (the
eft,
wha
atwastaken, p
permissiongive
en,
retu
urncar)
Theeft of other tha
an car or moto
or
veh
hicle (theft, wh
hatwastaken) ffrom
hom
me or near home (location__1) of
lesss than $10.
Theeft of other tha
an car or moto
or
veh
hicle (theft, wh
hatwastaken) ffrom
hom
me or near home (location__1) of
$10
0-$49
Theeft of other tha
an car or moto
or
veh
hicle (theft, wh
hatwastaken) ffrom
hom
me or near home (location__1) of
$50
0-$249
Theeft of other tha
an car or moto
or
veh
hicle (theft, wh
hatwastaken) ffrom
hom
me or near home (location__1)
valu
ue wasn’t give
en.
Theeft of other tha
an car or moto
or
veh
hicle (theft, wh
hatwastaken) ffrom
hom
me or near home (location__1) of
$25
50 or greater
Theeft of other tha
an car or moto
or
veh
hicle (theft, wh
hatwastaken) of
lesss than $10.
Theeft of other tha
an car or moto
or
veh
hicle (theft, wh
hatwastaken) of
0-$49
$10
Appe
endix F
Type of C
Crime Specifica
ations
Situatio
on
TOC=00 and
a SC731a=1
1 and (SC748=
=14,7-27 or SC748=yes ‘N
NA’) and
sum<250
TOC reccode
TOC=00 and
a SC731a=1
1 and (SC748=
=14,7-27 or SC748=yes ‘N
NA’) and
sum=99998
28
TOC=00 and
a SC731a=1
1 and (SC748=
=14,7-27 or SC748=yes ‘N
NA’)
27
TOC=00 and
a any SC642
2_04-05=1
90
NCVS-CS Pilot
P
Report
26
F-5
TO
OC
descrription
Compl eted
Person
nal
Larcen
ny
withou
ut
Contacct ($50$249)
Compl eted
Person
nal
Larcen
ny
withou
ut
Contacct
(Value NA)
Compl eted
Person
nal
Larcen
ny
withou
ut
Contacct
($250+
+)
Unwan
nted
Sexuall
Contacct
Recode d
description
Theeft of other tha
an car or moto
or
veh
hicle (theft, wh
hatwastaken) of
$50
0-$249
Theeft of other tha
an car or moto
or
veh
hicle (theft, wh
hatwastaken) vvalue
wassn’t given.
Theeft of other tha
an car or moto
or
veh
hicle (theft, wh
hatwastaken) of
$25
50 or greater
wanted sexuall contact with or
Unw
with
hout force (wh
hathappen)
Appendix G
Description of the Type of Crime Codes
Appen
ndix G
Codes
Descriiption of the TType of Crime C
Description of Type of Crime co
odes (crimess listed in oorder of seve
erity)
TOC
code
01
TOC descripttion
Completed Rape
R
02
Attempted Rape
03
Sexual
S
Attack
k with
Serious Assa
ault
Respondentt is present (Q
Q20B=1 or 2)
&
Attack/Injury/
A
/Threat = sexu
ual assault (Q2
29=3 or Q31=
=3 or Q28=5,6
6)
&
{ A Weapon is present
p
(Q23 n
not missing) *
*OR* there is a serious inju
ury
(Q31=5 thru
u 9) *OR* the re is a minor injury (Q31=1
10-11) with a
hosp
pitalization lassting more tha
an a day (Q37
7>1) }
04
Sexual
S
Attack
k with
Minor Assault
05
Completed
d
Robbery
R
with Injury
I
from Seriou
us
Assault
Respondentt is present (Q
Q20B=1 or 2)
&
{ Atta
ack/Injury/Threat = sexual assault (Q29=
=3 or Q31=3 o
or Q28=5,6) w
with a
Mino
or injury (Q31=
=10,11) *OR*
* there is unw
wanted sexual contact (Q28=
=5,6)
with a minor threatt such as bein
ng followed or surrounded (Q
Q28=12) or th
here is
an
n attempt/threeat to hit, slap
p etc. (Q28=13
3) }
Respondentt is present (Q
Q20B=1 or 2)
&
ng was stolen
Somethin
n (Q88=1)
&
{ Th
here is a seriou
us non-sexual injury (Q31=5
5-9, stab wounds thru knoccked
unconscious)
u
*OR*
*
a minorr injury (Q31=10, bruises) o
or “other” injurry
(Q3
31=11) with a Weapon pressent (Q23=1-6
6) *OR* there
e is some “oth
her”
injury (Q31=11)
(
with
h a 2+ day hosspital stay (Q3
37>1) }
08
Attempted
A
Rob
bbery
with Injury frrom
Serious Assa
ault
Respondentt is present (Q
Q20B=1 or 2)
&
There was an
n attempt to ssteal (Q89=1)
&
{ Th
here is a seriou
us non-sexual injury (Q31=5
5-9, stab wounds thru knoccked
unconscious)
u
*OR*
*
a minorr injury (Q31=10, bruises) o
or “other” injurry
(Q3
31=11) with a Weapon pressent (Q23=1-6
6) *OR* there
e is some “oth
her”
injury (Q31=11)
(
with
h a 2+ day hosspital stay (Q3
37>1) }
11
Aggravated
A
Asssault
Completed with
w
Injury
Respondentt is present (Q
Q20B=1 or 2)
&
{ Th
here is a seriou
us non-sexual injury (Q31=5
5-9, stab wounds thru knoccked
unconscious)
u
*OR*
*
a minorr injury (Q31=10, bruises) o
or “other” injurry
(Q3
31=11) with a Weapon pressent (Q23=1-6
6) *OR* there
e is some “oth
her”
injury (Q31=11)
(
with
h a 2+ day hosspital stay (Q3
37>1) }
NCVS-CS Pilot
P
Report
Variables used to deterrmine TOC
Respondentt is present (Q
Q20B=1 or 2)
&
Attack/Injuryy = rape (Q29=
=1 or Q31=2)
Respondentt is present (Q
Q20B=1 or 2)
&
{ Atta
ack/Injury = atttempted rapee (Q29=2 or Q
Q31=3) *OR* there was a TThreat
off rape (Q28=1
1) with a weap
pon present (Q
Q28=7) or the R was struck or
surround
ded (Q28=8 th
hru 13) }
G-1
Appen
ndix G
Codes
Descriiption of the TType of Crime C
TOC
code
06
TOC descripttion
Completed
d
Robbery
R
with Injury
I
from
f
Minor Asssault
Variables used to deterrmine TOC
Respondentt is present (Q
Q20B=1 or 2)
&
ng was stolen
Somethin
n (Q88=1)
&
There
T
is a min
nor injury (Q31
1=10, bruisess) or “other” injjury (Q31=11)
09
Attempted
A
Rob
bbery
with Injury frrom
Minor Assault
Respondentt is present (Q
Q20B=1 or 2)
&
There was an
n attempt to ssteal (Q89=1)
&
There
T
is a min
nor injury (Q31
1=10, bruisess) or “other” injjury (Q31=11)
14
Simple Assa
ault
Completed with
w
Injury
Respondentt is present (Q
Q20B=1 or 2)
&
There
T
is a min
nor injury (Q31
1=10, bruisess) or “other” injjury (Q31=11)
07
Completed
d
Robbery with
hout
Injury
Respondentt is present (Q
Q20B=1 or 2)
&
ng was stolen
Somethin
n (Q88=1)
&
{ The
ere was a non--sexual attackk (Q29=4-11, sshot thru attempted attack
k with
oth
her weapon) *OR* a threat//attempt at a non-sexual atttack (Q28=2--14,
threat/atttempt to kill th
hru “other” thrreat/attempt to attack }
10
Attempted
A
Rob
bbery
without Inju
ury
Respondentt is present (Q
Q20B=1 or 2)
&
There was an
n attempt to ssteal (Q89=1)
&
{ The
ere was a non--sexual attackk (Q29=4-11, sshot thru attempted attack
k with
oth
her weapon) *OR* a threat//attempt at a non-sexual atttack (Q28=2--14,
threat/atttempt to kill th
hru “other” thrreat/attempt to attack }
12
Attempted
d
Aggravated
A
Asssault
with Weapo
on
Respondentt is present (Q
Q20B=1 or 2)
&
{ The
ere was a non--sexual attackk (Q29=4-11, sshot thru attempted attack
k with
oth
her weapon) *OR* a threat//attempt at a non-sexual atttack (Q28=2--14,
threat/atttempt to kill th
hru “other” thrreat/attempt to attack }
&
There was a W
Weapon preseent (Q23=1-6))
13
Threatened
T
Asssault
with Weapo
on
Respondentt is present (Q
Q20B=1 or 2)
&
There was a W
Weapon preseent (Q23=1-6))
15
Sexual Assa
ault
without Inju
ury
Respondentt is present (Q
Q20B=1 or 2)
&
Attacck/Threat = seexual assault (Q29=3 or Q2
28=4)
NCVS-CS Pilot
P
Report
G-2
Appen
ndix G
Codes
Descriiption of the TType of Crime C
TOC
code
16
TOC descripttion
Unwanted Se
exual
Contact with
hout
Force
Variables used to deterrmine TOC
Respondentt is present (Q
Q20B=1 or 2)
&
Threat = unwanted seexual contact w
without force (Q28=6)
17
Assault without
Weapon with
hout
Injury
Respondentt is present (Q
Q20B=1 or 2)
&
{ Minor attack (Q29=10,12-1
14) *OR* Threeat/Attempt tto Kill or Attacck
(Q
Q28=2,3) paire
ed with a Threeat/Attempt to follow/surro
ound or hit/sla
ap
(Q28
8=12,13) *OR
R* Threat/Atteempt of a minor attack (Q28
8=11,13-14 w
which
includes atttempt/threat oof thrown objeect, hit/slap/e
etc., and the
“oother/specify””) }
18
Verbal Threa
at of
Rape
Respondentt is present (Q
Q20B=1 or 2)
&
Threat/Atttempt = Rapee (Q28=1)
19
Verbal Threa
at of
Sexual Assa
ault
Respondentt is present (Q
Q20B=1 or 2)
&
Threat/Attemp
T
pt = Sexual asssault (Q28=4
4)
20
Verbal Threa
at of
Assault
Respondentt is present (Q
Q20B=1 or 2)
&
Threat/Attemp
T
pt to Kill or Atttack (Q28=2,3
3)
21
Completed Purse
Snatching
g
Respondentt is present (Q
Q20B=1 or 2)
&
A purrse was stolen
n (Q96=2) whiich was on thee victim’s “perrson” at the tim
me of
the theft (Q102A = 1)
23
Pocket Picking
(Completed only)
o
Respondentt is present (Q
Q20B=1 or 2)
&
Ca
ash (Q96=1) or
o a wallet (Q9
96=3) was sto
olen which wass on the victim
m’s
“person” at the time of the th
heft (Q102A = 1)
22
Attempted Pu
urse
Snatching
g
Respondentt is present (Q
Q20B=1 or 2)
&
Offfender attempted to steal a purse (Q90=
=2) which wass on the victim
m’s
“perrson” at the tim
me of the atteempt (Q102A = 1)
31
Completed
d
Burglary, Forccible
Entry
Attempted
A
Forrcible
Entry
Completed
d
Burglary, Unla
awful
Entry
E
Without Force
Attempted Motor
Vehicle The
eft
onse)
Therre is evidence that offenderr got in using fforce (Q15 hass a ‘yes’ respo
an
nd the offendeer actually gott inside (Q12=
=1)
33
32
41
NCVS-CS Pilot
P
Report
Therre is evidence that offenderr got in using fforce (Q15 hass a ‘yes’ respo
onse)
The offender actually got in
nside (Q12=1))
There is an attempted
a
thefft (Q89=1) of a motor vehiccle (Q90=5,6)
G-3
Appen
ndix G
Codes
Descriiption of the TType of Crime C
TOC
code
39
29
40
34
TOC descripttion
Attempted
d
Household
H
Larrceny
Variables used to deterrmine TOC
e is an attemp
pted theft (Q89
9=1) from thee HH primary p
property (Q10
0=1,2)
There
or near theeir property (Q
Q10=5,6,7)
Attempted
d
Personal Larcceny
without Conttact
Completed Motor
M
Vehicle The
eft
There is an attempted th
heft (Q89=1)
e is a theft (Q8
88=1) of a mootor vehicle (Q9
96=5,6) and tthe offender d
did not
There
have the right to borrow
b
the car (Q98 ne 1) orr the offenderr had permissiion to
borrow
b
but did not return thee car (Q99 = 2
2)
Completed
d
Household
H
Larrceny
(less than $1
10)
Completed
d
Household
H
Larrceny
($10-$49)
e is a theft (Q8
88=1) from th e HH primary property (Q10
0=1,2) or nearr their
There
prop
perty (Q10=5,6
6,7) and the ccost of the stollen items sum
ms to between
n $10
and $49
36
Completed
d
Household
H
Larrceny
($50-$249
9)
e is a theft (Q8
88=1) from th e HH primary property (Q10
0=1,2) or nearr their
There
prop
perty (Q10=5,6
6,7) and the ccost of the stollen items sum
ms to between
n $50
and $249
38
Completed
d
Household
H
Larrceny
(Value NA
A)
Completed
d
Household
H
Larrceny
($250+)
e is a theft (Q8
88=1) from th e HH primary property (Q10
0=1,2) or nearr their
There
property (Q1
10=5,6,7) and the cost of th
he stolen itemss is unknown
Completed
d
Personal Larcceny
without Conttact
(less than $1
10)
Completed
d
Personal Larcceny
without Conttact
($10-$49)
Completed
d
Personal Larcceny
without Conttact
($50-$249
9)
Completed
d
Personal Larcceny
without Conttact
(Value NA
A)
Completed
d
Personal Larcceny
without Conttact
($250+)
Q88=1) and th
he cost of the stolen items ssums to less tthan
There is a theft (Q
$10
35
37
24
25
26
28
27
NCVS-CS Pilot
P
Report
e is a theft (Q8
88=1) from th e HH primary property (Q10
0=1,2) or nearr their
There
property (Q10=5,6
6,7) and the coost of the stollen items sum
ms to less than
n $10
There
e is a theft (Q8
88=1) from th e HH primary property (Q10
0=1,2) or nearr their
property (Q10=5,6
6,7) and the coost of the stolen items sum
ms to $250 or more
ere is a theft (Q
Q88=1) and th
he cost of the stolen items sums to betw
ween
The
$10 and $49
9
ere is a theft (Q
Q88=1) and th
he cost of the stolen items sums to betw
ween
The
$
$50 and $249
9
There
T
is a thefft (Q88=1) and
d the cost of tthe stolen item
ms is unknown
n
There
e is a theft (Q8
88=1) and thee cost of the sstolen items su
ums $250 or more
G-4
Appen
ndix G
Codes
Descriiption of the TType of Crime C
TOC
code
90*
TOC descripttion
Unwanted Se
exual
Contact
Variables used to deterrmine TOC
The
ere was unwan
nted sexual co
ontact (Q27=4
4,5)
NOTE: Varia
able numbers are
a based on the
e 2006 NCVS2 (Westat CATI vaariables may bee named differe
ently).
* TOC code 90 is excluded from BJS NCVS
S published estimates of crimee.
NCVS-CS Pilot
P
Report
G-5
Appendix H
Census Instructions for Using the Automated NCVS
Editing and Coding Blaise Instrument
ndix H
Appen
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
Census Instructio
PROJECT 7523008
7523008-9009-C
SECTION
N1
100-30-09
NATIONA
AL CRIME VIC
CTIMIZATIO
ON SURVEY (N
NCVS)
SE
ECTION 1
INSTRUCT
TIONS FOR USING
U
THE A
AUTOMATED NCVS
EDITIN
NG AND COD
DING BLAISE
E INSTRUMEN
NT
TABLE OF CONTEN
NTS
Page
Section
A.
B.
C.
D.
E.
F.
G.
H.
I.
J.
K.
L.
M.
ODUCTION.................................................................................................................... 2
INTRO
MATERIALS NEED
DED ......................................................................................................... 3
GENE
ERAL INFORM
MATION ................................................................................................ 3
GENE
ERAL EDIT IN
NSTRUCTION
NS .................................................................................... 4
HOW TO
T RECODE THE ENTRY
Y IN RACE-SPE
ECIFY ............................................... 6
HOW TO
T RECLASSIIFY AN INCID
DENT........................................................................... 8
NSTRUMENT
SCREE
ENS IN THE EDITING
E
AND CODING IN
T................................. 9
HOW TO
T EDIT INC
CIDENTS FLA
AGGED FOR D
DELETION IIN
PRODUCTION .......................................................................................................................17
HOW TO
T REFER IN
NCIDENTS..........................................................................................20
HOW TO
T PROCESSS DUPLICATE
ES ..................................................................................22
HOW TO
T DELETE INCIDENTS ......................................................................................24
HOW TO
T HANDLE
E INCIDENTS FLAGGED F
FOR DELETIO
ON IN E/C ...........25
EXITING A CA
ASE..................................................................................................................................27
ATTACHM
MENTS
A
B
C
List of Crime Reclassification Codes and List of Refferral Reason C
Codes
Race In
nstructions for Control Card screen RACE
Weapon
ns Instructions For Crime Inccident Report sccreen WEAPO
ON
H-1
NC
CVS-CS Pilot R
Report
Appen
ndix H
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
100-30-09
NATIONA
AL CRIME VIC
CTIMIZATIO
ON SURVEY (N
NCVS)
SE
ECTION 1
INSTRUCT
TIONS FOR USING
U
THE A
AUTOMATED NCVS
EDITIN
NG AND COD
DING BLAISE
E INSTRUMEN
NT
A.
IN
NTRODUCTIO
ON
1
The Editingg and Coding (E
E/C) instrumen
nt is an automatted system thatt replaces the paaper and pencill
method of correcting
c
NCV
VS protocols. However,
H
the saame general rulles for correctin
ng paper protoccols
apply to editting/coding casses with the E/
/C instrument.
This proced
dure is divided into
i
two section
ns. Section 1 pprovides generall information o
on the E/C process
and Section 2 provides detailed E/C instrructions.
2
3
The Automated System peerforms the follo
owing functionns:
a
The ediiting and codin
ng of the form in
n one pass.
b
Allows adding, changiing, or deleting of data as neceessary.
c
Automatically skips to
o the correct quuestion after ediiting the data annd blanks out aall entries that aare no
longer relevant.
r
d
Electro
onically refers caases to Headquuarters (HQ) forr resolution.
This section
n of the proceduure describes th
he basic functioons of the Editiing/Coding instrument.
a
Accessiing the Editing/Coding instruument.
b
Retrievving an NCVS CAPI
C
case to bee edited/codedd.
c
Movingg through the Editing/Coding
E
g instrument.
d
Editingg and coding off duplicate incid
dents.
e
How to
o refer an incideent.
f
How to
o delete and und
delete incidentss.
NCVS-CS Pilot
P
Report
H-2
ndix H
Appen
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
g
B.
100-30-09
Exitingg the instrument at the middle of coding an “IIncident” and eexiting at the en
nd of a “Case”.
MATERIALS
M
NEEDED
N
1
Access to th
he E/C instrum
ment for NCVS CAPI cases.
2
A copy of th
he NCVS WebC
CATI Coding Navigation
N
Mannual, Proceduree Number 75233084-603-C.
3
A copy of th
he “Instruction
ns For Using thee Automated N
NCVS Editing aand Coding Blaaise Instrument,,”
Procedure Number
N
7523-X
XXX.
4
List of Crim
me Reclassificatiion Codes and List
L of Referrall Reasons (Attaachment A)
5
Instructionss for Editing Co
ontrol Card screeen RACE (Atttachment B)
6
Instructionss for Editing Crrime Incident Report
R
screen W
WEAPON (Atttachment C)
7
Copy of thee NCVS-2 Items Booklet
8
Copy of thee latest NCVS-5550 CAPI Manuual for Field Reepresentatives
9
Copy of thee list of cases neeeding to be revviewed (printedd from WebCAT
TI)
100 Copy of thee list that detailss the reason(s) why
w a case is eliligible for E/C (included in th
he e-mail from H
HQ,
which initiattes the start of the NPC’s review of that monnth’s cases)
C.
GENERAL
G
INF
FORMATION
N
1
Approximattely 6,400 cases a month are in
nterviewed usinng the CAPI Blaaise Instrumentt. Of those,
approximateely 800 cases peer month will reequire editing aand coding. Thhe actual numbeer of cases requuiring
verification in the NPC maay vary over tim
me; some monthhs the NPC maay have more th
han or less than
n 800
cases to reviiew.
2
Editing and coding files aree generated forr each CAPI casse with: incidennt reports; entriies in screen(s)
SQCALLPOLICECRIM
ME; SQNOCAL
LLPOLICECR
RIME; or writte-in entries in tthe Control Carrd
screen RAC
CE.
3
All editing and
a coding is do
one electronicallly.
4
The NCVS CAPI cases aree transmitted to
o Jeffersonville for clerical editting and codingg on a monthly basis.
NCVS-CS Pilot
P
Report
H-3
Appen
ndix H
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
D.
100-30-09
5
Editing and coding begins approximately on the 4th of eaach month andd closes out app
proximately on tthe
26th of the same
s
month.
6
“F7 Item Leevel Notes” Pagge - Any notes that the Field R
Representative (FR) intervieweer makes duringg the
interview are displayed as “F7
“ Item Level Notes.” Thesee notes are linkked to the questtion where the FR
invoked the “F7 Notes” po
op-up window. A screen that contains a papper clip symbol next to the ansswer
category willl have an “F7 Item
I
Level Notte.” To view thhese notes deprress the “Shift” and “F7” keys
simultaneou
usly.
7
“CAPI Notees” - Any notess that the FR makes
m
after the iinterview, beforre exiting a casee. This is the fi
first
screen you will
w see once yo
ou enter a case in
i the WebCAT
TI system. Theese are notes an
n FR enters to ffurther
clarify someething from the interview or to
o make a note tto assist them w
with future interrviews with thee
household. This is also wh
here you enter notes
n
after editiing a case and bbefore exiting tthe case. Notice that
these “CAPI Notes” contaain notes from the
t FR since thhe case was firstt in sample, so some notes willl not
pertain to th
he month you are
a editing; theyy are from a preevious enumeraation period. T
The newest notees are
displayed at the bottom off the screen, wh
hich is the defauult view when yyou first view th
he “CAPI Notees.”
GENERAL
G
ED
DIT INSTRUC
CTIONS
1
Refer to Pro
ocedure Numbeer 7523084-6033-C for instructtions on accessiing and exiting the WebCATI
system and on retrieving caases to review and
a edit.
2
Navigate thrrough the instruument by hittin
ng “Enter” to ggo to the next qquestion. You ccan also use thee
arrow keys to
t go forward or
o backwards in
n the instrumennt.
3
Change an existing
e
value to
o a new value by
b either, deletinng the old valuee and typing in the new value, or by
simply typin
ng the new valuue over the existting value and hhitting “Enter.””
4
In the “Marrk All That App
ply” screens, sep
parate multiple precodes with commas or sep
parate the entries
with spaces..
5
Precodes caan also be selectted or deselecteed by using the mouse to clickk on the radio o
or square button
n to
the left of th
he precode.
6
If there is a write-in “Otheer-Specify” entrry, read the desccription displayyed in the speciffy line and try tto
reclassify th
he entry into on
ne of the existin
ng precoded caategories. If thee write-in entry is in a “Mark aall that
apply” screeen, be sure to deselect the preccode containingg the “Other-sppecify” in additiion to selectingg the
new precode.
NCVS-CS Pilot
P
Report
H-4
ndix H
Appen
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
100-30-09
7
An FR may enter a “Blind”” “Don’t Know
w” when there iis not a “Don’t Know” precodde. This is enteered by
simultaneou
usly hitting the “Control”
“
+ “D
D” keys. It willl appear as a yeellow question m
mark in the E/C
C
instrument.
8
An FR may enter a “Blind”” “Refusal” when a respondennt refuses to answer a question
n. This is enterred by
simultaneou
usly hitting the “Control”
“
+ “R
R” keys. It will appear as a bluue exclamation mark in the E//C
instrument.
9
Do not edit crimes for houusehold membeers who are Nooninterviews. T
They will have a status of “DO
ONENonint” under the “STA” column in screeen SCREEN11. Only edit crim
me reports for household mem
mbers
who have a status of “DON
NE-Int.”
100 Edit only th
he items that aree specified to be edited.
111 Duplicate reeporting of inciidents - If theree is any indicatioon that more thhan one househ
hold member iss
reporting th
he same inciden
nt, or the FR has flagged the caase as a duplicaate, refer the inccidents to HQ. (See
Section 1, Su
ubsection J.)
122 Do not refer the followingg types of incideents unless you are unable to rresolve them ussing the specificc
instructions given later in this
t procedure.
a
Arson (The
(
malicious act of burning a respondent's home or any bbuilding on the respondent's
propertty.)
b
Fraud only
o (For examp
ple: Embezzlem
ment, con gam
mes, extortion, fo
forged checks, illlegal use of creedit
cards.)
c
Peepingg Toms/Exhibiitionism only.
d
Telepho
one threats, thrreatening letterss, e-mail, or phoone texts only.
e
Cases which
w
were deleeted because the reference perriod is out-of-sccope.
133 Utilize the “CAPI
“
Notes” to
t fill in any miissing data. Skiip patterns are eedited by the co
omputer. How
wever,
if a skip patttern is changed
d during editing, the screens thhat are missing ddata (in the new
w skip pattern) must
be filled clerrically, when ap
pplicable, based
d on the SUMM
MARY screen aand Notes section.
144 Enter any notes or questions you have about a case in thhe “CAPI Notees” before exitin
ng a case.
155 Follow the detailed
d
E/C in
nstructions as outlined in Sectiion 2 of this proocedure.
NCVS-CS Pilot
P
Report
H-5
Appen
ndix H
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
100-30-09
16
6 Contact Katthryn Cheza, in
n Crime Survey
ys Branch (CS
SB), Demograpphic Surveys D
Division (DSD), at
HQ by e-maail when you have
h
completed
d reviewing/verrifying all of thhe CAPI cases.
E.
HOW
H
TO REC
CODE THE EN
NTRY IN RAC
CE-SPECIFY
Y
1
If precode (6),
( “Other-speccify” is marked for Race, the R
RACEROSTE
ER_EC screen will appear. It
follows the SCREEN1 scrreen OR the “C
Catchall” questiion(s) screen(s) (if there are an
ny) and precedees the
incident rep
port.
2
The RACEROSTER_EC
C screen displayys: the line num
mber (“LNO”), the respondentt’s name (“NAM
ME”),
their relation
nship to the refference person (“REL”), their Hispanic originn (“ORG”), thee selected race
precode (“R
RACE”), and th
he “Other-speciify” race (“RAC
CE SPECIFY”)) under the headers in parenth
hesis.
This is show
wn in the screen
n shot illustrated in Figure 1.
3
At the requeest of our spon
nsor, the Bureauu of Justice Stattistics (BJS), all write-in entriess of race must b
be
reviewed, when
w
“Other-speecify”, Precodee (6) is selected,, in order to dettermine if the ““other race” can
n be
classified intto one of the fiive precoded raace categories. T
To accommodaate this requestt, the E/C instrrument
will display the
t screen RAC
CEROSTER_
_EC when Preccode (6), “Otherr-specify” was selected duringg the
interview in the RACE scrreen. It is this information
i
youu will review annd attempt to edit into one of the
five precodeed race categoriies.
NCVS-CS Pilot
P
Report
H-6
ndix H
Appen
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
4
100-30-09
Attempt to reclassify the “O
Other-specify” race entries, ussing the instrucctions found in Attachment B.. In
Figure 1 thee “Other-specify
fy” entry of “Hiispanic” shouldd be reclassifiedd to Precode (1)), “White.”
Figure 1
1
John Boe
2
Mary Boe
3
Ivan Boe
4
Barry Boe
Barry Boe
Ivaan Boe
5
The instrum
ment cycles thro
ough the Race and
a “Catchall” sscreens for eachh household m
member before ggoing
to the incideent report(s). For
F example, if LN1 reports ann “Other-speciffy” race entry, tthe instrument will
show the racce entry first an
nd then progresss to any “Catchhall” questions for LN1. Nexxt the instrumen
nt will
go to any “O
Other-specify” race entries forr LN2 and thenn to any “Catchhall” questions ffor LN2 and
continue in this order for all
a respondents in the househoold.
Therefore, after
a
editing thee race code for a respondent thhe instrument w
will progress th
hrough any otheer
“Other-speccify” race entriees and “Catchalll” questions foor the householdd in the followiing order:
a
To CAT
TCHALL1_EC or CATCHA
ALL2_EC if thhis respondent also reported aan incident at on
ne or
both off the crime screeener “Catchall”” screens (See SSection 1, Subsection F for insstructions on how to
reclassify an incident reported
r
in the crime screenerr “Catchall” screeens.)
OR
NCVS-CS Pilot
P
Report
H-7
Appen
ndix H
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
Census Instructio
PROJECT 7523008
7523008-9009-C
SECTION
N1
b
100-30-09
To RACEROSTER_
_EC if the nextt person on thee roster reportedd an “Other-sp
pecify” race
OR
c
To onee of the “Catchaall” screens (CA
ATCHALL1_E
EC or CATCH
HALL2_EC) iff the next perso
on on
the rostter reported an incident at onee or both of thee crime screeneer “Catchall” sccreens, but did n
not
report an
a “Other-speccify” race.
OTHE
ERWISE,
d
F.
The insstrument will prrogress to the INCROSTER_
I
R_EC screen aftter you have fin
nished
reviewing/coding all of
o the RACEROSTER_EC, CATCHCALL
L1_EC and CA
ATCHALL2_E
EC
screenss.
HOW
H
TO REC
CLASSIFY AN
N INCIDENT
1
If the respon
ndent answered
d “Yes” to Item
m SQCALLPO
OLICECRIME
E and/or Item
SQNOCAL
LLPOLICECR
RIME, these “Catchall” questtion screens willl automaticallyy come up after
SCREEN1, before any inccidents. There are two correspponding “Catchhall” screens, C
CATCHALL1__EC
and CATCH
HALL2_EC.
NCVS-CS Pilot
P
Report
H-8
Appen
ndix H
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
2
100-30-09
In Figure 2 the Incident Description of “N
Noisy neighborrs” has been coorrectly classifieed as “Not a Crrime”,
Precode (277).
Figure 2
Joh
hn Boe
Maary Boe
Ivaan Boe
G.
3
Use the “F11” key to accesss the Help Screeen for coding thhe “Catchall” sscreens.
4
Identify up to three crime classification
c
co
odes that best ddescribe what hhappened, basedd on the inciden
nt
description.
5
After classiffying the inciden
nt, continue co
oding the remainning “Catchall”” and/or Race sscreens for the
household.
6
The instrum
ment cycles thro
ough the Race and
a “Catchall” sscreens for eachh household m
member before ggoing
to the incideent report(s). Refer
R
to Section
n 1, Subsection E5 for more innformation on h
how the instrum
ment
cycles throu
ugh these three screens.
`S
SCREENS IN
N THE EDITIN
NG AND COD
DING INSTRU
UMENT
NCVS-CS Pilot
P
Report
H-9
Appen
ndix H
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
Census Instructio
1
T_EC is the inttroductory screeen in the E/C instrument. Pllease refer to th
he NCVS WebC
CATI
The START
Coding Navvigation Manuall, Procedure Nuumber 75230844-603-C for a sccreen shot exam
mple. You will enter
“1” at this screen to contin
nue on in the in
nstrument.
NCVS-CS Pilot
P
Report
H-10
ndix H
Appen
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
2
100-30-09
SCREEN1 is the next screeen in the E/C instrument. A
As shown in Figgure 3, this screen lists all of th
he
people on th
he household ro
oster and their interview and hhousehold statuus. This includes:
a
“LNO”” - The line num
mber of each peerson on the rooster.
b
“NAME
E” - Name of each
e
person on
n the roster by liine number.
c
“SLF/P
PRXY” - Indicaates whether th
he respondent’s interview was completed as a “Self” intervieew or
“Proxy” interview.
d
“STA” - Status of theiir interview, wh
hether an intervview, noninterviiew or not eligiible because thee
person is under 12 or a nonmember.
e
“AGE”” - Age of each line number.
f.
“REL” - Relationship of each line nuumber to the reeference personn.
g
“SEX” - Gender of eaach line numberr.
h
“MEM
M” - Whether or not each line number
n
is a mem
mber of the hoousehold.
i
“DATE
E OF LAST IN
NTERVIEW” - The last time eeach line numbber was interview
wed (current
intervieew or previous interview
i
date, when noninterrview this enum
meration).
NCVS-CS Pilot
P
Report
H-11
Appen
ndix H
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
j
100-30-09
“HHR”” - The Househ
hold Responden
nt is indicated bby an “X”.
Figure 3
1117 CHURCH
ANYTOWN,
A
CA
C 99997
1
John Boe
2
Mary Boe
3
Ivan Boe
4
B
3
B
The INCRO
OSTER_EC screen comes neext and lists all of the incidentts reported for tthe household. This
screen appears after the SC
CREEN1 screeen when no onee in the househhold has an “Otther-specify” raace nor
either of thee “Catchall” scrreens. Otherwiise, INCROST
TER_EC appeaars after codingg the
RACEROS
STER_EC, thee CATCHALL
L1_EC and/or tthe CATCHAL
ALL2_EC screeens. It containss the
following in
nformation:
a
The nu
umber of inciden
nts for the houusehold are listeed sequentially iin the first coluumn on the left.. This
is also the
t number youu will enter to access
a
an incideent.
b
The “L
LNO” column is the line numb
ber of the respoondent who repported the inciddent.
c
The “IN
NC” column is the incident nuumber. The inccidents are num
mbered sequenttially at the persson
level. That
T means thaat the first incid
dent report for eeach line numbber will be “1” iin this column aand
the seco
ond incident reeport for each liine number willl be “2”, etc. T
This is illustrated in Figure 4.
d
The “ST
TA” column diisplays the statuus of the incideent.
NCVS-CS Pilot
P
Report
H-12
Appen
ndix H
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
100-30-09
There are
a four possiblle “Status” codees, which are auutomatically uppdated in the instrument when
n an
inciden
nt is edited:
(1))
“P” means the
t incident is Pending
P
and nneeds to be reviiewed/edited.
(2))
“R” means the incident haas been Referreed to HQ.
(3))
“D” means the incident haas been Deleted
d.
(4))
“C” means the incident haas been Compleeted. It has beeen reviewed an
nd verified with
hout
any referralss.
e
The “R
REF CODE” co
olumn displays the Referral Coode(s). The coddes in this coluumn describe w
why the
inciden
nt was referred to
t HQ. This co
olumn is autom
matically updated in the instrum
ment when an
inciden
nt is referred.
f
The “W
WHAT HAPPE
ENED” column
n gives a brief ddescription of eeach incident.
In Figu
ure 4 there are three incidents reported
r
for the household. L
Line 1 reportedd incidents 1 andd 2;
Line 2 reported
r
incideent 3. Incidentss 1, 2 and 3 are pending review
w.
Figure 4
NCVS-CS Pilot
P
Report
H-13
Appen
ndix H
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
100-30-09
4
n incident to reeview/edit, enteer the number ffrom the far lefft column that corresponds w
with
To select an
the incidentt you want to review.
5
After you haave completed reviewing/editiing an incident,, the instrumennt will return to this
INCROST
TER_EC screen
n.
6
If you have completed reviiewing/editing all of the incideents for a case (or after codingg all Race and
“Catchall” screens
s
in a casee without incideents), enter “311” to exit the caase. The instruument will proceeed to
the WHATSTAGE_EC screen,
s
which iss covered in Secction 1, Subsecttion G7 below..
7
The WHAT
TSTAGE_EC screen appears when “31” is eentered at the IINCROSTER__EC screen to exit
the case. Th
his screen displlays which stagee of review the case is currentlly in. There aree three stages o
of
review that a case goes thro
ough during thee editing and cooding: “Initial,” “Verification
n,” and “Fina.l.”” The
first two staages are done att the NPC, whille the “Final” SStage of review is done at HQ.. Only cases in
n
which a refeerral reason hass been entered will
w get to the thhird stage. Othherwise, a case will close-out aafter
the “Verification” Stage.
In Figure 5 below, the casse is currently in
i the “Initial” Review stage of E/C.
Figure 5
NCVS-CS Pilot
P
Report
H-14
Appen
ndix H
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
a
b
100-30-09
The WH
HATSTAGE_E
EC screen askss if you are donne completing thhis stage of the review.
(1)
Iff you answer Prrecode (1), “Yess, done reviewiing this case,” yyou will be direccted to the pop
p-up
window
w
AREYO
OUSURE_CK
K (Figure 6). Thhis screen is covvered in Section
n 1, Subsection
n G7b
beelow.
(2)
Iff you answer Prrecode (2), “No
ot done reviewinng this case,” yyou will exit thee case. The casee will
reemain in the currrent stage of review.
The AR
REYOUSURE
E_CK pop-up window
w
will apppear next if youu answered “Yees, done with
reviewing this case” in
n screen WHAT
TSTAGE_EC , as explained bbelow.
(1)
The AREYOUS
SURE_CK screeen asks “Are yyou sure you are ddone with (Initial//Verification/Finnal)
Review?”
R
The insstrument will auutomatically filll in the name of the stage you are currently in
n, in
th
he pop-up wind
dow.
(2)
As
A Figure 6 illusttrates, if you arre done press thhe “Suppress” bbutton to contin
nue exiting the case.
Iff you are not do
one, press the “Close”
“
or “Gooto” button to rreturn to the W
WHATSTAGE__EC
sccreen and continue your review
w.
NCVS-CS Pilot
P
Report
H-15
Appen
ndix H
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
(3)
100-30-09
Iff you press the “Suppress” buttton the instrum
ment will updatte the current sttage of review ssince
yo
ou said you werre done. Thereefore, the “Supppress” button w
will advance thee status of the ccase
frrom “Initial” to
o “Verification”” or from “Veri fication” to “Fiinal” when refeerral codes havee been
en
ntered in either the “Initial” orr “Verification”” Stages. Otherrwise, if no refeerral codes were
en
ntered the statuus of the case will
w advance from
m “Verificationn” to “Completted”.
Figu
ure 6
NCVS-CS Pilot
P
Report
H-16
Appen
ndix H
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
8
100-30-09
The SHO_F
FINAL screen
n is the last screeen you will see before exiting the E/C instruument. From th
his
screen you will
w return to th
he WebCATI syystem to enter yyour notes in thhe “CAPI Notees” screen befo
ore
exiting the case.
c
As illustraated in Figure 7,
7 enter “1” to ccontinue from tthis screen. Th
he SHO_FINA
AL
screen also shows
s
the finall control codes that are passedd onto WebCAT
TI for the case.
Figure 7
H.
NCIDENTS FLA
AGGED FOR DELETION
D
IN P
PRODUCTION
HOW TO EDIT IN
1
If an inciden
nt has been flaggged for deletio
on by the FR, thhe incident will have a status o
of “D” in the sccreen
INCROST
TER_EC when the case is firstt accessed. Thee screen WHY
YDELETEDIN
NPROD_EC w
will
appear afterr entering an inccident number to review in thee INCROSTE
ER_EC screen,, in which an FR
R had
flagged the incident
i
for delletion.
Incidents may
m be flagged for
fo deletion by the
t FR in produuction for the ffollowing reasons:
NCVS-CS Pilot
P
Report
H-17
Appen
ndix H
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
100-30-09
a
“Out of Scope” - The incident happeened outside off the current refference period. If the incident is
deleted because it is “O
Out of Scope,”” a description oof the reason it was deleted wiill be displayed,, such
as “Hap
ppened outsidee of the referencce period” (Seee Figure 8). If aan incident was deleted by the FR,
becausee the incident happened
h
outsid
de the referencee period, you caan delete the in
ncident once yo
ou
reach th
he INCIDENT
TDATE screen
n. For instructtions on deletinng incidents, seee Section 1,
Subsecttion K.
b
“Dupliccate” - The inciident is a dupliccate of anotherr incident reporrted in the curreent enumeration
n
period or it is a duplicate of an incideent reported in a prior enumerration period. IIf the incident iis
flagged as a “Duplicatte”, the two digiits in the “INC
C NBR” columnn will be “96-999”. In the screeen
shot sh
hown in Figure 9, the incident is a duplicate oof an incident reeported in the ccurrent enumerration,
becausee the incident number
n
is not “96-99”. Insteaad, it is “02,” meaning it was co
oded as a dupliicate
of the second
s
incidentt reported durin
ng the current eenumeration. N
Note that the in
nterview numbeer
(“INT NBR”)
N
of “2” and
a Line Numb
ber of the persoon reporting thhe other inciden
nt (“01”) are alsso
displayeed. For more in
nformation on handling dupliicates see Sectioon 1, Subsection
n J.
Figure 8
NCVS-CS Pilot
P
Report
H-18
Appen
ndix H
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
100-30-09
Figure 9
2
DELINCR
ROSTER_EC Screen
a
If a case has been flagged for deletion
n by the FR thee incident was eeither coded as “happening ouutside
the refeerence period” or as a “duplicaate,” as discusseed in Section 1, Subsection H1 above.
b
The scrreen DELINC
CROSTER_EC
C will appear affter reviewing thhe reason why an incident wass
flagged for deletion att the WHYDEL
LETEDINPR
ROD_EC screeen.
c
The inccident being flaggged for deletio
on will have a SStatus Code of “D” in the “ST
TA” column.
d
Below (Figure 10) is an
n example of an
n incident that was flagged forr deletion as dissplayed at the
DELIN
NCROSTER_
_EC screen. Th
his screen is onn path after the WHYDELET
TEDINPROD
D_EC
screen (Figures 8 and 9) or the screen
n WHYDELE
ETEDINEC_E
EC (Figure 14)..
NCVS-CS Pilot
P
Report
H-19
Appen
ndix H
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
100-30-09
Figure 10
0
e
Inciden
nts that have beeen deleted in errror may be unndeleted after ree-accessing the
DELIN
NCROSTER_
_EC screen. To
o undelete the iincident, enter Precode (1). Iff you choose to
o
undelette an incident, the
t instrument will
w automaticaally advance to tthe SUMMAR
RY_EC
scrreen,
in order to re-edit the entire incident again.
Note:
I.
C using this screeen. These incidents
You currentlly cannot undellete an incident deleted in E/C
can only be undeleted
u
by restarting the casse.
f
To leavve the incident deleted,
d
enter Precode
P
(2). Thhe instrument aautomatically reeturns to the
INCRO
OSTER_EC screen. At this point,
p
continuee coding incidennts or exit the ccase.
g
To leavve the incident deleted
d
and revview the data, ennter the Precodde (3). The insttrument will
automaatically advance to the SUMM
MARY_EC screeen, in order to review or re-eddit the entire in
ncident
again.
HOW
H
TO REFER INCIDENTS
1
Refer an inccident to HQ att any time whilee editing an inciident.
NCVS-CS Pilot
P
Report
H-20
Appen
ndix H
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
100-30-09
2
To refer an incident to HQ
Q, simultaneoussly hit the “Conntrol” and “W” keys from with
hin the incidentt
report. Pleaase note that it does not matteer what screen yyou are on, as loong as you are iin the incident
report.
3
The followin
ng pop-up screeen will appear, listing all of th e possible referrral reasons.
Figure 11
1
4
Select the reeferral code num
mber(s) that best describes whhy the incident iis being referreed. Attachmentt A
also contain
ns this list of refferral codes. In
n Figure 11 the incident is beinng referred for Referral Reason
n (32),
“Theft & Atttempted Theftt Problems.”
5
Up to six referral reasons may
m be marked per incident. IIf there are morre than six enteered, a pop-up eerror
message will appear, sayingg to limit the nuumber of codess to six or less.
6
After enterin
ng in the referrral reason(s), cliick on the “OK
K” button on thhe pop-up screeen. The instrum
ment
automaticallly returns to thee screen in the incident reportt where you invvoked the referrral code pop-up
p
window.
7
o edit the rest off the incident reeport for otherr possible reasons to refer the incident as welll as to
Continue to
edit items needing to be recoded. To refeer the incident ffor another reaason(s), repeat ssteps 2 through
h 6.
NCVS-CS Pilot
P
Report
H-21
Appen
ndix H
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
8
To delete an
n incident, simuultaneously hit the
t “Control” aand “W” keys aand select Referrral Code (40),
“Delete Inciident” from thee pop-up screen
n. Then enter tthe reason the iincident is bein
ng deleted. See
Section 1, Su
ubsection K fo
or more informaation about deleeting incidents..
Note:
J.
100-30-09
If an incident is deleted
d
in error,, the case will hhave to be restarrted in order to
o undelete the
inccident.
HOW
H
TO PRO
OCESS DUPL
LICATES
1
Duplicate in
ncidents, which were flagged as
a duplicates by the FR, will haave a Status Code of “D” in th
he
“STA” column of the INC
CROSTER_EC
C screen.
2
Refer to Secction 1 Subsectiion H1a(2) for more informatition on identifyi
ying incidents co
oded as duplicaates by
the FR.
3
Any inciden
nts flagged as a duplicate in pro
oduction by thee FR are alwayss referred to HQ
Q.
NCVS-CS Pilot
P
Report
H-22
Appen
ndix H
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
4
100-30-09
Review the descriptions for each incident under the "WH
HAT HAPPEN
NED" column iin the
INCROST
TER_EC to ideentify possible duplicates
d
not aalready flagged by the FR in prroduction, as sh
hown
in the screen
n shot below (F
Figure 12). Notte that in this exxample, there aare two incidents of a bike theft
reported by LN 2.
Figure 12
1
a
If any of
o the summariees look similar, the incidents m
may be duplicaates.
b
Edit su
uspected duplicaate incidents firrst.
c
Review
w the suspected duplicate incidents to determiine if they actuaally are duplicattes.
(1)
Enter
E
the inciden
nt number of th
he first possiblee duplicate inciddent to review at the
IN
NCROSTER_
_EC screen in order
o
to review
w it.
(2)
Review
R
the first possible dupliccate incident, paaying particularr attention to th
he SUMMARY
Y
sccreen and any Notes.
N
(3)
Enter
E
the inciden
nt number of th
he second posssible duplicate inncident at the IINCROSTER
R_EC
sccreen.
NCVS-CS Pilot
P
Report
H-23
ndix H
Appen
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
(4)
d
(2)
K.
Review
R
the secon
nd suspected duuplicate incidennt the same wayy, paying particcular attention tto the
SU
UMMARY scrreen and any Notes.
N
Use thee information from
fr
the SUMM
MARY screen aand Notes as w
well as all pertin
nent data (such aas the
date off the incident an
nd number of tiimes), to determ
mine if the inciddents are dupliccates.
(1)
e
100-30-09
Iff the incidents are
a determined to be duplicatees, refer both/aall of the inciden
nts to HQ as
du
uplicates, follow
wing the instrucctions in Sectioon 1, Subsectionn I, “How to Refer Incidents.””
(aa)
Enter the Referral
R
Code (19) for the susppected duplicatte(s) and contin
nue editing and
coding the incident for otther possible reaasons to refer tthe incident to H
HQ.
(b
b)
“Reaccess,”” “Recode,” an
nd “Refer” the ffirst possible duuplicate after co
oding the secon
nd
incident as a possible duplicate. Continuue to review thiis incident for o
other possible
reasons to refer the incideent to HQ.
Iff the second (an
nd any subsequent) possible duuplicate is not a duplicate, con
ntinue editing aand
co
oding the incideent using standard E/C proceedure.
Althouggh it is rare, households may have
h
triplicate ((or more) incideents or two or m
more sets of
duplicaate incidents. Enter
E
the inciden
nt number of aany subsequent suspected dupllicate incidents and
repeat the
t same steps listed above.
HOW
H
TO DELETE INCIDEN
NTS
1
When flaggiing an incident for deletion, sim
multaneously hhit the “Controll” + “W” keys from within the
incident rep
port. Please notte that it does not
n matter whatt screen you aree on, as long as you are in the
incident rep
port. (Also see Section 1, Subssection I, “How
w to Refer Inciddents”.)
NCVS-CS Pilot
P
Report
H-24
Appen
ndix H
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
2
100-30-09
The followin
ng pop-up screeen will appear, listing all of th e possible referrral reasons.
Figure 13
1
L.
3
Select Preco
ode (40), “Delette Incident”. A write-in box w
will appear. Ennter the reason w
why the inciden
nt is
being deleteed. In Figure 133, the incident is
i being deletedd because it “haappened outsidee of the referen
nce
period.” Affter entering thee reason click on
o the “OK” buutton and then hit “Enter” to advance to thee
WHYDELETEDINEC_
_EC screen.
4
From the WHYDELETE
W
EDINEC_EC screen enter “11” to proceed too the DELINC
CROSTER_E
EC
screen. Thee incident beingg deleted will haave a Status Coode of “D” in thhe “STA” colum
mn. Enter Preccode
(2) to leave as is and return
n to the INCRO
OSTER_EC s creen. (The DELINCROST
TER_EC screeen is
discussed in
n more detail in Section 1, Subsection H2.)
LETION IN E
HOW
H
TO HAN
NDLE INCIDE
ENTS FLAGG
GED FOR DEL
E/C
1
An incident may either be flagged for deleetion in producction by the FR
R (“Out-of-Scop
pe” or “Duplicaate”),
or an incident may be flaggged for deletion
n in E/C.
NCVS-CS Pilot
P
Report
H-25
ndix H
Appen
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
Census Instructio
PROJECT 7523008
7523008-9009-C
SECTION
N1
100-30-09
a
If an in
ncident was flaggged for deletion
n by the FR, thhe incident beinng flagged for ddeletion will havve a
Status Code
C
of “D” in
n the “STA” collumn of the IN
NCROSTER_E
EC screen, andd will not have a
referrall reason in the “REF
“
CODE” column. See SSection 1, Subseection H, “How
w to Edit Incideents
Flagged
d for Deletion in
i Production” for more detaills on incidents flagged for deletion by the FR
R.
b
If an in
ncident was flaggged for deletion
n in E/C the inncident being fllagged for delettion will have a status
of “D” in the “STA” column
c
and a code
c
of “40” in the “REF COD
DE” column in
n the
INCRO
OSTER_EC screen.
2
If you selectt an incident at the INCROST
TER_EC screeen that has beenn flagged for deletion in E/C the
instrument will
w proceed to the WHYDEL
LETEDINEC
C_EC screen.
3
This screen,, as shown belo
ow (Figure 14), displays the reaason why the inncident was flaggged for deletio
on by a
coder who previously
p
revieewed the incideent. In this exam
mple the incideent was flagged for deletion beecause
it is “not an NCVS crime.””
Figure 14
4
Enter “1” to
o continue. Th
he instrument will
w proceed to tthe DELINCR
ROSTER_EC screen, which iis
covered in Section
S
1, Subseection H2.
NCVS-CS Pilot
P
Report
H-26
Appen
ndix H
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
5
If you agreee with the reason the incident was
w flagged forr deletion in E//C, select Preco
ode (2) at the
DELINCR
ROSTER_EC screen. Otherw
wise, select Preccode (3) at the DELINCROS
STER_EC scrreen
and once in the incident reefer to HQ for resolution.
r
Note:
M.
100-30-09
Baased on how thee instrument is currently set-upp, once an inciddent is flagged for deletion in E/C
it cannot
c
be unflaagged within thee instrument. T
Therefore, Preccode (1) to “Un
ndelete and reeddit” is
no
ot functioning properly.
p
Refer these incidentss to HQ.
EXITING A CA
ASE
1
You may exxit an incident/ccase at any poin
nt by pressing tthe “F10” key. Although this functionality exxists, it
is recommen
nded that you only
o exit using the
t instructionss in step 2, beloow.
2
When you have
h
completed
d your review off the case, enterr “31” at the IN
NCROSTER__EC screen. Affter
entering “311” in the INCR
ROSTER_EC screen, the insttrument will prroceed to the sccreens
WHATSTA
AGE_EC, ARE
EYOUSURE_
_CK, and SHO
O_FINAL. These screens are discussed in Seection
1, Subsectio
on G, “Screens in
i the Editing and
a Coding Insstrument.”
3
If you need to exit a case prior
p
to finishing your review eenter Precode ((2) at the WHA
ATSTAGE_EC
C
screen in order to keep thee case in the currrent review staage.
4
For
Once you have progressed past these screeens you will exxit the instrumeent and return tto WebCATI. F
W
refeer to the “NCV
VS WebCATI C
Coding Navigatiion
instructions on exiting out of the case in WebCATI,
Manual,” staarting on page 1-12.
NCVS-CS Pilot
P
Report
H-27
ndix H
Appen
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N2
10-30-09
NATIONA
AL CRIME VIC
CTIMIZATIO
ON SURVEY (N
NCVS)
SE
ECTION 2
INSTRUCT
TIONS FOR USING
U
THE A
AUTOMATED NCVS
EDITIN
NG AND COD
DING BLAISE
E INSTRUMEN
NT
TABLE OF CONTEN
NTS
Section
A.
B.
C.
D.
E.
F.
G.
Page
ODUCTION.................................................................................................................... 1
INTRO
GENE
ERAL EDIT IN
NSTRUCTION
NS .................................................................................... 1
REFER
RRALS .............................................................................................................................. 2
SPECIF
FIC EDIT INSSTRUCTIONSS ...................................................................................... 3
HOW TO
T EDIT SPE
ECIFIC ITEMSS OF THE INC
CIDENT REPO
ORT.......................... 8
VERIF
FICATION .....................................................................................................................23
DISPO
OSITION ........................................................................................................................24
NCVS-CS Pilot
P
Report
H-28
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N2
10-30-09
NATIONA
AL CRIME VIC
CTIMIZATIO
ON SURVEY (N
NCVS)
SE
ECTION 2
INSTRUCT
TIONS FOR USING
U
THE A
AUTOMATED NCVS
EDITIN
NG AND COD
DING BLAISE
E INSTRUMEN
NT
A.
IN
NTRODUCTIO
ON
This section provvides instructio
ons for Editing and Coding (E
E/C) NCVS casses with inciden
nt reports, writee-in
en
ntries in the “Catchall” screenss SQCALLPO
OLICECRIME
E and SQNOC
CALLPOLICE
ECRIME, and w
writein
n entries in the Control
C
Card sccreen RACE. Refer unresolvved cases to DSD
D/CSB at Heaadquarters (HQ
Q) for
reesolution.
B.
GENERAL
G
ED
DIT INSTRUC
CTIONS
1
Perform
m the editing an
nd coding in on
ne pass.
2
Do nott edit crimes forr household meembers who haave a status of ““DONE-Nonin
nt” under the “SSTA”
column
n in the SCREE
EN1 screen. Only
O edit crime reports for houusehold membeers who have a status
of “DO
ONE-Int."
3
Edit on
nly the screens specified.
s
Sincee the Blaise insttrument edits th
the data, it is no
ot necessary to eedit
the following:
a
Skip paatterns - are editted by the computer. Howeveer, if a skip patttern is changedd during editingg, the
screenss that are missin
ng data (in the new
n skip pattern
rn) must be filleed clerically, wh
hen applicable, b
based
on the Summary and Notes
N
sections.
b
omputer. Therrefore, fractionss and ranges are not clerically
Numerric entries - are edited by the co
edited.
4
Refer to
o Section 1, Sub
bsection D of this
t procedure oon how to makke corrections tto existing data..
5
Contact Kathryn Chezza, DSD/CSB, via email whenn you have com
mpleted reviewin
ng all of the CA
API
cases.
NCVS-CS Pilot
P
Report
I-1
endix I
Appe
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N2
C.
10-30-09
REFERRALS
R
1
Once a casee is accessed, revview and edit (iif necessary) anny household m
members who haave a write-in eentry
in screen RA
ACE_SPECIF
FY and reclassiffy any incidentss based on entriies to screens
SQCALLPOLICECRIM
ME and SQNOCALLPOLICE
ECRIME (Seee Section 1, Sub
bsections E andd F,
respectively)). Refer to Attaachment A for the List of Crim
me Reclassificaation Codes andd Attachment B for
Race Instrucctions for Conttrol Card Screen
n RACE for m
more details on rreviewing and eediting race.
2
Edit all incid
dents in all casees. After a referral reason is foound, continue editing the inciident in order tto
resolve or fo
or any additional reasons referr to HQ.
3
Resolve inciidents that repo
ort Commerciall Crimes as listeed below. (For example: burglary or larceny o
of
respondent's place of emplloyment.)
a
If theree is any indicatio
on an Incident Report is for a commercial crrime, determinee if the crime in
ncident
occurreed against a reco
ognizable or an
n unrecognizablle business:
1)
Un
nrecognizable business
b
- If youu determine theere was a crimee against an unrecognizable buusiness,
acccept the NCVSS incident reporrt. For examplee, the respondeent runs a busin
ness from homee, and
theere is no sign on
n the premises advertising thee business. This could be men
ntioned in the N
Notes
or SUMMARY screen.
s
If you are
a unable to deetermine if the business run frrom home is an
n
un
nrecognizable buusiness, refer th
he incident to H
HQ.
2) Reco
ognizable busin
ness:
a)
If you determ
mine there was only a crime aggainst the recoggnizable businesss and no otherr
NCVS crimee occurred then
n delete the inciddent as “Out-oof-Scope." (Forr instance, theree was
no illegal enttry or attempted
d illegal entry oof the respondennt's home, etc.;; no attack or
attempted atttack occurred; no
n one was thrreatened with hharm in any wayy; or no househ
hold
property wass stolen.)
b) If you determ
mine there weree personal itemss stolen from a recognizable b
business, edit sccreens
THEFT thrrough RECOV
VEREDINSUR
RANCE on thhe NCVS incideent report to acccept
only the resp
pondent's perso
onal loss (not thhe business losss). For examplee, if the responddent's
sweater was stolen at the offfice, or if the reespondent's lunnch was stolen from the refriggerator
at work, the sweater and lun
nch would be thhe respondent'ss personal loss. If unable to reesolve,
refer the inciident to HQ.
NCVS-CS Pilot
P
Report
I-2
endix I
Appe
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N2
b
10-30-09
If you are
a unable to deetermine wheth
her the crime occcurred in a reccognizable or un
nrecognizable
businesss, refer the inciident to HQ.
4
If an inciden
nt involves a po
olice officer as the
t offender, reefer the incidennt to HQ.
5
Hit and run accidents:
6
a
If theree is NO indicatiion there was a DELIBERAT
TE attempt to ruun over the victim or run the vvictim
off the road (not just an accident or poor driving), ddelete the inciddent as “Out-off-Scope."
b
If theree is an indication of DELIBER
RATE intent too run over the rrespondent or rrun the responddent
off the road refer the incident
i
to HQ
Q. (For examplee, a situation wh
where the respon
ndent knew thee
offendeer).
c
If in do
oubt about the intent,
i
refer thee incident to HQ
Q.
Duplicate reeporting of inciidents:
If there is an
ny indication th
hat more than one
o household m
member is repoorting the samee incident, or th
he
Field Repressentative (FR) has
h flagged the case as a dupliccate, refer the iincidents to HQ
Q. (See Section
n 1,
Subsection J.)
J
7
If an inciden
nt report is inco
omplete it is pro
obably becausee the person is a Type Z, Noniinterview. Theese
incidents aree to be referred
d to HQ using Referral
R
Code 33.
8
Do not refer the followingg types of incideents unless you are unable to rresolve them ussing the specificc
instructions given later in this
t procedure.
a
Arson - The maliciouss act of burningg a respondent'ss home or any bbuilding on thee respondent's
propertty.
b
Fraud only
o (For examp
ple: Embezzlem
ment, con gamees, extortion, foorged checks, illlegal use of credit
cards.)
c
Peepingg Toms/Exhibiitionism only.
d
Telepho
one threats, thrreatening letterss, E-mail, or phhone texts only..
e
Cases which
w
were deleeted because the reference perriod is “Out-of--Scope."
D. SPEC
CIFIC EDIT IN
NSTRUCTION
NS
1
Recoding raace write-in entrries - See Attach
hment B for m
more details.
NCVS-CS Pilot
P
Report
I-3
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N2
2
10-30-09
Examine all FR Notes and “F7 Item Leveel Notes," then edit as followss:
a
“Speciffy Screens” - Determine if entrries for “Speciffy” screens can be reclassified into a precodedd
categorry.
1) Pay close attention
n to the followiing screens, ass the entries in these screens ccan
usuaally be reclassiified into a preccoded categoryy:
LOCATION
N_SPEC
HOW
WATTACK_SPEC
WHATHAP
PPEN_SPEC
ATT
TEMPTTHEF
FTWHAT_SPE
EC
HOWATTA
ACK_SPEC
WHATWAS
STAKEN_SPE
EC
HOWTHRE
EATEN_SPE
EC
DOINGATINCIDENTT
TIME_SPEC
2) If ab
ble to reclassifyy a specified enttry in a multiplee answer categoory, delete the
“Specify” precode and select the correct
c
precodee. If you reclasssify a “Specify””
answ
wer on a single answer categorry, select the coorrect precode. You do not neeed to delete th
he
“Specify” precode for single entryy questions.
NO
OTE:
3)
b
Do not
n reclassify screens
s
EMPL
LOYERTYPE
E through
JOBMSATYPE.
Fo
or “Specify” enttries in the “Cattch-All” screenns SQCALLPO
OLICECRIME
E or
SQ
QNOCALLPO
OLICECRIME
E reclassify anyy incidents usingg the instruction
ns in Section 1,,
Subsection F. If there are more than three inciidents displayedd in the “Incideent Description
n” in
thee screen(s) CAT
TCHALL1_EC
C and/or CAT
TCHALL2_EC
C reclassify onlyy the first three
inccidents.
Summaary, Notes - Determine if the Summary
S
and N
Notes conflict w
with any of the data entries.
1)
If a conflict existss with an NCVSS screen listed iin Section 2, Suubsection E of this procedure,,
w
the incideent report to aggree with the Suummary and N
Notes. If you aree
corrrect the data within
un
nable to resolve based on the Summary,
S
refer to HQ.
2)
If a conflict existss with an NCVSS screen not lissted in Section 22, Subsection E of this proceddure,
corrrect the data within
w
the incideent report to aggree with the Suummary and N
Notes. If you aree
un
nable to resolve based on the Summary,
S
acceppt the related inncident report eentries; do not rrefer
theese incidents.
3)
If unable
u
to resolvve because the Summary or N
Notes are vague,, refer to HQ.
NCVS-CS Pilot
P
Report
I-4
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N2
3
10-30-09
Edit the typ
pes of incidents shown below as
a follows:
a
Police officer
o
as victim
m - Accept the incident
i
and eddit as usual. (Coonsider any law
w enforcement o
officer
a policee officer.)
b
Propertty Belonging to
o Nonhousehold
d Member - If the Notes or Suummary indicaate that the inciddent
only invvolves propertyy that belonged
d to a nonhouseehold member oor a recognizab
ble business, sch
hool,
etc., change the entry in
i screen THE
EFT or ATTEM
MPTTHEFT to “2." If in ddoubt or unablee to
resolve, refer to HQ. (See Section 2, Subsection E ffor more detaills on coding thee theft and attem
mpted
theft qu
uestions for this scenario.)
c
Theft/A
Attempted Theeft of Householld Member's Prroperty from N
Nonhousehold M
Member or
Recogn
nizable Businesss - If the incideent involves houusehold properrty that was loan
ned to or left w
with a
nonhou
usehold membeer or a recognizzable business, sschool, etc., edi
dit as follows:
1) If th
he property wass stolen from th
he nonhousehoold member, buusiness, school,
etc., accept the inciident.
2) If th
he nonhouseholld member, bussiness, school, eetc., kept the prroperty, sold it,,
desttroyed it, etc., but
b the propertyy was not stolenn from the nonnhousehold
mem
mber, business, school, etc., ch
hange the entry in screen THE
EFT or
ATT
TEMPTTHEF
FT to “2."
3) If in
n doubt, refer to
o HQ.
d
Stolen Antenna
A
- If an
n incident involvves a stolen mootor vehicle anttenna, edit as fo
ollows:
1) Accept the incident if the incidentt reports a theft
ft and there are no Notes to
conttradict it.
2) Accept the incident if the incidentt does not repoort a theft and tthere are no No
otes
to co
ontradict it.
3) Chaange screen TH
HEFT to (1) “Y
Yes” or (2) “Noo” to agree withh the Notes or tthe
Sum
mmary when possible.
4) If yo
ou are unsure as to whether th
he antenna was stolen or not, rrefer to HQ.
e
Arson, Fraud, Peepingg Tom/Exhibittionism, Telephhone, E-mail, Phhone Text, or L
Letter Threats, and
Indirect Threats Relayyed to Respond
dent by Personss Other Than thhe Offender -
NCVS-CS Pilot
P
Report
I-5
endix I
Appe
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N2
1)
10-30-09
Th
hese incidents are
a not NCVS crimes
c
unless annother NCVS ccrime happenedd in conjunction
n with
theem. Do not reffer these incidents unless you are unable to reesolve them usiing the followin
ng
insstructions:
a)
Review the in
ncident and thee Summary to ddetermine if anoother NCVS
crime also occcurred, such ass:
1.
Was theere an illegal enttry or attemptedd illegal entry? (Does someth
hing in the
Summarry/Notes indicaate an illegal enntry or attempteed illegal entry, or is Precode ““11”
selected for screen LOCATION_GE
ENERAL, or is Precode “16”” and/or Precodde
“17” sellected in screen
n WHATHAPP
PEN?)
2.
Was thee respondent atttacked, was theere an attemptedd attack, or wass the responden
nt
threaten
ned with harm in any way? (D
Does something in the Summarry/Notes indicaate an
attempteed or actual attaack or threat, oor is Precode “115” or “16” seleected in screen
WHATHAPPEN or are
a any precodees selected in sccreens HOWT
TRYATTACK
through HOWATTAC
CK?)
3.
Was theere household property
p
stolen or an attempt tto steal househo
old property? ((Does
something in the Summ
mary/Notes inddicate theft or aattempted theftt, or are any preecodes
selected in screens ATT
TEMPTTHE
EFTWHAT or WHATWAST
TAKEN?)
N
crime did occur in conjuunction with a nnon-NCVS crim
me (arson, fraudd,
b) If another NCVS
Peeping Tom
m/exhibitionism
m, telephone, e--mail, phone texxt, or letter threats; indirect th
hreats
relayed to thee respondent by persons otherr than the offennder), delete the non-NCVS ccrime
(arson, fraud
d, Peeping Tom
m/exhibitionism
m, telephone, e-m
mail, phone texxt, letter threatss, or
indirect threaats) from the in
ncident reports.
c)
2)
If no other NCVS
N
crime occcurred in conjuunction with thhe non-NCVS ccrime (arson, fraaud,
Peeping Tom
m/exhibitionism
m, telephone, e--mail, phone texxt, or letter threats; indirect th
hreats
relayed to resspondent by peersons other thaan the offenderr), delete the inccident as “Out--ofScope.”
Exxamples of each
h of the previouusly listed non-N
NCVS crimes ooccurring in
con
njunction with another NCVSS crime are giveen below.
a)
NCVS-CS Pilot
P
Report
Arson - The offender stole the respondentt's car and thenn burned the resspondent's garaage.
Delete the arrson from the report
r
and keepp the theft of thhe car. If the offfender had onlly
burned the reespondent's garrage delete the incident as “Ouut-of-Scope."
I-6
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N2
10-30-09
Note:
Arsson is the maliccious act of burrning the respoondent's housee or any buildinng on
the respondent's property.
p
b) Fraud - The respondent's puurse containingg a wallet and crredit cards was stolen and the
offender used the credit carrd to buy merchhandise. Keep the theft of thee purse, wallet, and
credit cards. However, if th
he incident incluudes the value oof the merchan
ndise that was
purchased fraudulently, deleete the amount of the merchanndise from the value in screen
n
PROPERTY
YVALUE. If only
o the credit ccard number waas used to buy merchandise an
nd no
theft had occcurred (the cred
dit card was nott stolen), deletee the incident ass “Out-of-Scop
pe.”
Note:
c)
Exaamples of fraud arre embezzlement, con games, extorttion, forged checkss, illegal use of creddit
cardd, theft of electricity
ty, theft of water frfrom outside faucet,t, etc.
Telephone, Letter,
L
E-mail, or
o Phone Text Threat - The ooffender telepho
oned/sent a lettter, email, etc. to the
t respondent and threatenedd to beat up thee respondent. T
The offender th
hen
came to the respondent's
r
ho
ome and illegallly entered the rrespondent's ho
ome and threateened
to kill the resspondent. Deleete the telephonne, letter, e-maiil, etc. threat to beat up the
respondent and
a keep the illeegal entry and tthe threat to killl (since this thrreat was made ffaceto-face). If there
t
was only a telephone or lletter threat andd no face-to-facce confrontatio
on had
occurred, dellete the inciden
nt as “Out-of-Sccope.”
d) Peeping Tom
m/Exhibitionism
m - The responndent saw the ooffender lookingg through her
window late at night. She screamed and thhe offender expposed himself th
hen fled. As hee left,
the offender stole an ornam
mental statue thaat was sitting inn the responden
nt's yard. Delette any
entries related to looking th
hrough the winddow, exposing hhimself, or the trespassing. K
Keep
the theft of the
t statue. If on
nly the incidentt of a Peeping T
Tom and the exxhibitionism haad
occurred and
d no theft had occurred,
o
deletee the incident as “Out-of-Scop
pe.”
e)
Indirect Threeat - The respondent was shoppping and ran iinto a friend wh
ho told the
respondent that
t a person th
hey both knew hhad threatened to harm the reespondent. Afteer the
friend left, th
he respondent noticed
n
that herr wallet was gonne. Keep the th
heft of the walllet, but
delete the thrreat. If the walllet had not beeen stolen and onnly the indirectt threat was madde,
delete the inccident as “Out--of-Scope.”
3) If in
n doubt about any
a of these typ
pes of incidentts, refer the inccident to HQ.
NCVS-CS Pilot
P
Report
I-7
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N2
4
E.
10-30-09
f
Vandaliism - If an incid
dent involves only
o vandalism (no other NCV
VS crime occurrred), delete the
inciden
nt as “Out-of-Sccope."
g
Kidnap
pping - Is not an
n NCVS crime.. For incidents that report kiddnapping, refer to HQ.
h
Cases which
w
were deleeted because the incident occuurred outside thhe reference perriod - Accept ass is;
do not refer. This is further
f
describeed in Section 1, Subsection H.
i
If in do
oubt about any incident, refer to
t HQ.
Blind “Don
n’t Know” Entriies:
a
If theree is a blind “Do
on't Know” (“C
Control” + “D”” keys) entry and a “Don't Kno
ow” precode is
provideed, select the “D
Don't Know” precode.
p
b
For scrreen LOCATIO
ON_GENERA
AL on the Inciddent Report, acccept a blind “D
DK” entry. If
Precode (36), “Somew
where else” is seelected and the specify entry iss “Don’t Know
w” or “Refused”” enter
a blind “Don’t Know”” or a blind “Reefused” in the L
LOCATION__GENERAL screen.
HOW
H
TO EDIT
T SPECIFIC ITEMS OF TH
HE INCIDENT
T REPORT
IN
NCIDENTDA
ATE (Month and Year Incident Happened)
H
If month en
ntered is not witthin the referen
nce period and the FR coded tthe incident mo
onth as outside the
reference peeriod, delete thee incident usingg Referral Codee 40. (There wiill be a reason th
he incident wass
deleted in th
he screen WHY
YDELETEDINPROD_EC..)
However, iff the month app
pears to be insid
de the referenc e period, but thhe Notes and/o
or Summary inddicate
the incidentt actually happened outside thee reference periiod, enter Referrral Code 18 an
nd refer to HQ..
Example: The
T Reference Period
P
is Januarry 2009 - July 22009 and the FR
R entered March (Precode “3””) in
the screen INCIDENTDA
ATE, but the Summary
S
and/ or Notes mentition that the inccident happenedd in
March 20088. Refer to HQ using Referral Code 18.
NCVS-CS Pilot
P
Report
I-8
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N2
10-30-09
RECALLDETA
R
AILS (Recall Deetails)
If Precode (2)
( “No," “is a series”
s
is selected in screen RE
ECALLDETA
AILS enter Refferral Code 28 aand
refer to HQ
Q.
If the Summ
mary and/or No
otes indicate thiis incident is paart of a “Series”” enter Referrall Code 28.
IN
NCIDENTAIR (American Inddian Reservation)
If screen IN
NCIDENTAIR
R is blank, enteer Referral Codee 24 and refer tto HQ. Includee in the Notes
section in WebCATI
W
that INCIDENTA
I
AIR was blank.
LOCATION_G
GENERAL - LOCATION_
L
SPEC
(W
Where Did This Incident
I
Happen?)
?)
The location
n of where the incident took place
p
is coded uusing nine screeens. The first sccreen,
LOCATIO
ON_GENERA
AL, contains eigght “general” caategories that arre used to direcct the instrumen
nt to
one of the other
o
eight location screens to code a more sppecific locationn. The screens
LOCATIO
ON_IN_HOME, LOCATIO
ON_NEAR_H
HOME, LOCA
ATION_OTHE
ER_HOME,
LOCATIO
ON_COMMER
RCE, LOCATION_PARKIN
NG, LOCATIION_SCHOO
OL,
LOCATIO
ON_OPEN_AR
REA and LOC
CATION_SPE
EC are used to ccode the speciffic type of placee
where the in
ncident happened.
1
Examin
ne the Summaryy and Notes. Verify
V
that the eentry is correct.. If necessary, cchange
LOCA
ATION_GENE
ERAL to agreee with the Notees or Summary.
2
If you determine
d
that Precodes “11”--“14” should bee selected in LO
OCATION_IN
N_HOME tryy to fill
screenss OFFENDER
RLIVE through
h OFFENDE RGETIN by rreferring to the Summary or N
Notes
as necessary. If in douubt about how to code any of these screens, refer to HQ.
3
Try to resolve
r
those yo
ou are sure sho
ould be in Precoodes “15” throuugh “35” in scrreens
LOCA
ATION_NEAR
R_HOME thro
ough LOCATIION_OPEN__AREA (but aree not sure whicch
Precod
de “15”-“35”) based
b
on availab
ble informationn in the incidennt report or from
m the NCVS - 5550,
CAPI Interviewing
I
Maanual (pages B44-17 through B
B4-36). Use youur best judgmen
nt. If you are still
unable to resolve afterr reviewing the incident reportt and the manuual, select Referrral Code 2, andd refer
the inciident to HQ.
NCVS-CS Pilot
P
Report
I-9
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N2
10-30-09
4
S
does not
n indicate wh
here the incidennt took place, acccept the entry in the location
If the Summary
screenss as is.
5
If Preco
ode (36), “Som
mewhere else” iss selected in scrreen LOCATIO
ON_GENERA
AL reclassify th
he
location
n description in
n the screen LO
OCATION_SP
PEC into one oof the categories provided.
6
If “Don
n’t Know” is sp
pelled out or a blind
b
“DK” (“C
Control” + “D”” keys) is entered in
LOCA
ATION_SPEC enter a blind “DK”
“
in LOCA
ATION_GEN
NERAL. Similaarly, if “Refusedd” is
spelled out or a blind “RF” (“Contro
ol” + “R” keys)) is entered in L
LOCATION__SPEC enter a b
blind
“RF” in
n LOCATION
N_GENERAL
L.
7
If theree is any doubt about
a
the correcct category or aabout where thee incident happ
pened, enter Referral
Code 2, and refer to HQ.
H
EVIDENCE
E
(E
Evidence of Force or Attempted Forrce?)
If the only Precode
P
selected is “14," “18,"" or “19” and thhe write-in indiicates no force p
proceed as follo
ows:
1
Changee the entry in sccreen FORCED
DENTRY from
m “1” to “2."
2
Enter the appropriate precode(s) in screen
s
OFFEN
NDERGETIN
N.
OFFENDERG
O
GETIN (How Did
D Offender Get In/Try
I
To Get In
In?)
1
If Preco
ode “19," “Speecify” indicatess force, examinne the Summarry and Notes too determine if fforce
was acttually used.
2
If forcee was used:
a Chaange the entry in
n screen FORC
CEDENTRY from “2” to “1."
b Enteer the appropriate precode(s) in
i screen EVID
DENCE.
HHMEMBER
H
RPRESENT (H
Household Membeer Present When IIncident Occurredd?)
If the Summ
mary or Notes contradicts
c
the entry for HHM
MEMBERPRE
ESENT or WH
HICHMEMB
BER
take the folllowing action:
NCVS-CS Pilot
P
Report
I-10
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008 10-330-09
7523008-9009-C
SECTION
N2
1
n doubt about w
whether the resspondent or an
nother household
If the respondent was present, or if in
membeer was present, enter Referral Code
C
6 and refeer to HQ.
2
If accorrding to NCVSS rules (refer to the NCVS-5500 Manual, pagess B4-53 through
h B4-58) no
househ
hold member waas present, select Precode “2” for screen HH
HMEMBERPR
RESENT andd code
screenss KNOWLEAR
RNOFFENDE
ERS, SUREO
OFINFO and L
LEARNOFFE
ENDERS as
necessaary from inform
mation in the Suummary or from
m other coded sscreens.
3
If the answer
a
to screen
ns KNOWLEA
ARNOFFEND
DERS, SURE
EOFINFO or
LEARN
NOFFENDE
ERS is not know
wn, leave them blank. The resspondent wouldd have known tthe
answer to these questiions if screen HHMEMBER
H
RPRESENT w
was entered corrrectly during th
he
originall interview.
4
If theree is a discrepanccy between the first E/C revieew and the veriffication review enter Referral Code
6 and refer to HQ.
WHICHMEMB
W
BER (Which Members
M
Were Pressent?)
1
If the entry
e
in screen WHICHMEM
W
MBER is “3," bbut the Summarry/or Notes inddicate the respo
ondent
was preesent, enter Refferral Code 6 an
nd refer to HQ .
2
Try nott to refer if youur only question
n is whether thee respondent annd another household memberr were
presentt or only anotheer household member
m
was preesent (not the reespondent) in sscreen
WHIC
CHMEMBER. Use your judggment when revviewing these inncidents and co
ode the best youu can.
If you are
a unable to deetermine the rigght code, enter Referral Code 6 and refer to H
HQ.
3
If you determine
d
from
m the Summary and/or Notes that Precode “33” was marked correctly, accep
pt the
entry.
4
If Preco
ode “3” is seleccted in screen WHICHMEM
W
MBER, attempt to code screen
n HAPPEN fro
om
informaation in the Sum
mmary or from
m other coded sccreens.
5
If theree is a discrepanccy between the first E/C revieew and the veriffication review,, enter Referrall Code
6 and refer to HQ.
WEAPON
W
(Wass a Weapon Preseent/Type of Weappon?)
1
If Preco
ode “4," “5,” or “6” is selected
d and Precodess “1”-“3” are bllank, refer to A
Attachment C,
Instructions for WEA
APON. If unab
ble to resolve, eenter Referral C
Code 10 and reffer to HQ.
NCVS-CS Pilot
P
Report
I-11
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N2
2
10-30-09
ode (1) “Yes” is selected in scrreen WEAPON
NPRESENT and the weapon
n is selected in screen
If Preco
WEAP
PON, but Preco
ode (2) “No” is selected in scrreens ATTACK
K, TRYATTAC
CK and
THRE
EATEN enter Referral
R
Code 10
1 and refer to HQ.
WHATHAPPE
W
EN (What Actuaally Happened?)
1
When Precode
P
“20” iss selected in scrreen WHATHA
APPEN, if possible reclassifyy and edit as
necessaary. If unable to
o reclassify, revview the entry aand determine iif an NCVS crim
me has occurred. See
Section
n 2, Subsection D, pages 4 thro
ough 7 for desccriptions of whaat constitutes aan NCVS crimee.
a If an
n NCVS crime did occur, enteer Referral Codee 14 and refer tthe incident to
HQ.
b If no
o NCVS crime occurred, accep
pt the entry andd take no furthher action. The
incid
dent will not claassify.
2
If Preco
ode “13” is the only entry and
d the Summary does not indicaate something eelse happened, accept
the entrry as is; do not refer. If the Suummary indicattes another NC
CVS crime occuurred, enter Refferral
Code 14 and refer to HQ.
H See Sectio
on 2, Subsectionn D, pages 4 thhrough 7 for deescriptions of w
what
constitu
utes an NCVS crime.
c
3
If Preco
ode “14” or “155” is selected in
n screen WHAT
THAPPEN ennter Referral Code 7 and referr to
HQ.
4
If only Precode “18," and/or Precod
de “19” are seleccted take the foollowing action:
a Veriify that only Vaandalism occurrred (or was atteempted) and noo NCVS crime
occu
urred. (See Secction 2, Subsecttion D, pages 4 through 7 for ddescriptions off
whaat constitutes an
n NCVS crime.
b If on
nly Vandalism occurred
o
(or waas attempted), ddelete the inciddent as “Out-off
Scop
pe.”
HOWTRYATT
H
TACK, HOWT
THREATEN (Threatened/Trieed to Attack)
1
Reclasssify any “Specify
fy” entry in screeens HOWTRY
YATTACK_SP
PEC and
HOWT
THREATEN_
_SPEC, (Preco
ode “24”) if posssible. If unablle to reclassify, enter Referral C
Code
15 and refer to HQ.
NCVS-CS Pilot
P
Report
I-12
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N2
10-30-09
2
ode “11," “14,"" “15," or “16” is selected in sccreens HOWT
TRYATTACK or
If Preco
HOWT
THREATEN,, enter Referral Code 7 and reffer to HQ.
3
If only Precode “22” is
i selected in sccreens HOWTR
RYATTACK oor HOWTHR
REATEN, take the
followin
ng action:
a Examine the Summ
mary and determ
mine if anythingg else happenedd. If somethingg
else happened, cod
de accordingly. If in doubt as tto how to codee, refer to HQ.
b If yo
ou determine th
hat nothing elsee happened, acccept the incidennt as is and do
not refer. (If only Precode
P
“22” iss selected in scrreens HOWTR
RYATTACK
and HOWTHREA
ATEN take no
o further actionn. The incident will not classiffy.)
4
If the Summary
S
and/o
or any Notes cllearly indicates that an attack ooccurred, go baack to screen
ATTAC
CK and select Precode
P
“1." Select
S
all necessaary precodes inn screen HOWA
ATTACK. Usse the
Summaary, Notes and information
i
fro
om other screenns to complete screens PRET
THREATEN
through
h MEDICALE
EXPENSES, as
a necessary. Iff unable to fill oout, leave blankk.
5
If in do
oubt or unable to determine whether
w
it was aan attack, attem
mpted attack, oor threat, enter
Referraal Code 15 and
d refer to HQ.
HOWATTACK
H
K (Attacked)
1
If Preco
odes “11," “12”” and/or “13” is
i selected in H
HOWATTACK
K, enter Referraal Code 7 and reefer to
HQ.
2
If the Summary
S
and/o
or Notes clearlyy indicate(s) thhat an attemptedd attack or threeat occurred, go
o back
to screeen ATTACK an
nd change the Precode
P
from ““1” to “2." Theen select the necessary Precode for
screen TRYATTACK
T
K and/or THR
REATEN. Als o select the neccessary entries iin screen
HOWT
TRYATTACK
K or HOWTHR
REATEN.
3 If any dou
ubt or unable to
t determine whether
w
it was aan attack, attem
mpted attack, orr a threat, enterr
Referraal Code 31 and
d refer to HQ.
IN
NJURY (Injuriees)
1
If Preco
odes “12," “13,," and/or “14” are selected in the screen INJJURY, enter Reeferral Code 7 aand
refer to
o HQ.
NCVS-CS Pilot
P
Report
I-13
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N2
2
10-30-09
Try to reclassify
r
all entries in screen INJURY_SPE
I
EC to the precoodes in screen IINJURY. If
INJUR
RY_SPEC conttains an answerr that refers to aan emotional oor mental injuryy (i.e., not a phyysical
injury) delete Precode “21." If other injuries are maarked, keep those precodes. Iff no other injurries are
marked
d, select Precode “11." If unsuure, refer to HQ
Q.
HHMEMHAR
H
RMED (How Many
M Household Members
M
Harmedd?)
If there is an
n entry greater than or equal to
o “1," enter Reeferral Code 8 aand refer to HQ
Q.
HAPPEN
H
(Whaat Actually Happpened?)
If there is an
ny indication th
hat a household
d member was hharmed, threateened or robbedd by force, enterr
Referral Cod
de 14 and referr to HQ.
THEFT,
T
ATTE
EMPTTHEFT
T (Theft, Attemppted Theft)
1
If the Precode
P
in screeens THEFT or ATTEMPTT
THEFT is “1,"" examine the SSummary and N
Notes.
Verify that
t the precod
de is correct and
d change screenn THEFT or sscreen ATTEM
MPTTHEFT tto
agree with
w the Notes or
o Summary as necessary.
2
If the Precode
P
in screeen THEFT is “2”
“ and the Sum
mmary and/or Notes indicatee(s) that a houseehold
membeer's property waas taken, correcct the report as follows:
a Chaange screen TH
HEFT to “1."
b In sccreen WHATW
WASTAKEN add
a the stolen pproperty listed iin the Notes orr
Sum
mmary.
c Fill screens WHOO
OWNEDSTO
OLENPROPE
ERTY through
REC
COVEREDIN
NSURANCE using
u
informatiion in the Summ
mary and/or
Nottes.
d If in
n doubt, enter Referral
R
Code 32
3 and refer too HQ.
3
If screeen ATTEMPT
TEDTHEFT iss “2” and the SSummary and/oor Notes indicaate(s) someone
attemptted to take a ho
ousehold memb
ber's property, ccorrect the repoort as follows:
a Chaange screen ATT
TEMPTEDTHEFT to “1.""
NCVS-CS Pilot
P
Report
I-14
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N2
10-30-09
EMPTTHEFT
TWHAT througgh screen
b Fill screens ATTE
ATT
TEMPTTHEF
FTITEMS usiing informationn in the Notes aand/or Summaary.
c If in
n doubt, enter Referral
R
Code 32 and refer to H
HQ.
4
If THE
EFT is “1” and
d the Summary and/or Notes iindicate that soomeone attempted to take a
househ
hold member’s property,
p
correcct the report ass follows:
a Chaange screen TH
HEFT to “2” an
nd enter “1” in screen ATTEM
MPTHEFT.
b Use the Summary, Notes, and datta from screens WHATWAST
TAKEN througgh
ITE
EMSTAKEN to
t fill in screenss ATTEMPTT
THEFTWHAT
AT through
ATT
TEMPTTHEF
FTITEMS.
c If in
n doubt, enter Referral
R
Code 32 and refer to H
HQ.
5
If ATT
TEMPTTHEF
FT is “1” and th
he Summary annd/or Notes inddicate that a ho
ousehold memb
ber’s
propertty was taken, co
orrect the reporrt as follows:
a Chaange screen TH
HEFT to “1."
b Use the Summary, Notes, and datta from screens ATTEMPTT
THEFTWHAT
T
thro
ough ATTEMP
PTTHEFTITE
EMS to fill in W
WHATWAST
TAKEN througgh
REC
COVEREDIN
NSURANCE.
c If in
n doubt, enter Referral
R
Code 32 and refer to H
HQ.
ATTEMPTTH
A
HEFTWHAT (Attempted
(
Theft))
Examine thee Summary and
d Notes and verrify the correct categories werre selected. Chaange screen
ATTEMPT
TTHEFTWHA
AT to agree wiith the Notes oor Summary, as necessary.
1
If ATT
TEMPTHEFT
TWHAT is “266” review the Suummary and thhe “Specify” enttry for this screeen. If
possible, reclassify thiss entry to one of
o the precodedd categories.
2
If the Summary
S
and Notes
N
do not ind
dicate what thee offender tried to take, acceptt the entries.
3
If in do
oubt, enter Refe
ferral Code 32 and
a refer to HQ
Q.
NCVS-CS Pilot
P
Report
I-15
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N2
10-30-09
ATTEMPTTH
A
HEFTOWNER
R (Who Property Belongs To)
Examine thee Summary and
d Notes and verrify the correct categories werre selected. Corrrect as necessaary.
1
If the Summary
S
or No
otes indicate thaat the money orr property beloonged to a nonh
household mem
mber,
recogniizable business,, school, etc. (A
ATTEMPTTH
HEFTOWNER
R is “4”), to ressolve the incideent, do
the following:
a Reviiew the inciden
nt to determine if another NCV
VS crime occurrred. See
Secttion 2, subsectio
on D, pages 4 through
t
7 for ddescriptions of w
what constitutees
an NCVS
N
crime.
b If no
o other NCVS crime occurred
d, then delete thhe incident as ““Out-of-Scope.””
c If th
here is another NCVS
N
crime in
n conjunction w
with the attemppted theft of thee
nonhousehold prop
perty, delete thee attempted theeft of the nonhhousehold property
from
m screen ATTE
EMPTTHEFT
TWHAT. If n ecessary, edit sccreen
ATT
TEMPTTHEF
FT to Precode “2."
Exaample:
The respondent
r
rep
ports that the offfender attemptted to steal a frriend's
(non
nhousehold meember’s) motorrcycle from the respondent's ggarage during an
n illegal entry. D
Delete
the attempted thefft of the motorccycle by changinng ATTEMPT
TTHEFT to “2." Keep the
rem
maining part of the
t incident thaat includes the iillegal entry of rrespondent's ho
ome.
d If in
n doubt, enter Referral
R
Code 11
1 and refer to H
HQ.
2
If theree is any indicatio
on the money or
o property beloonged only to a household meember under 122, take
the following steps to correct the rep
port:
a Loo
ok at screens LO
OCATION_G
GENERAL, LO
OCATION_IN
N_HOME, an
nd
LOC
CATION_NE
EAR_HOME.
1)
If the precod
de in LOCATION_GENER
RAL is “11” or ““12," accept th
he incident as is.
Example:
A 9 year old household
h
mem
mber’s bike was stolen from th
he respondent’s front
yard (LOCA
ATION_GENE
ERAL is “12” and Precode “115” is selected iin screen
LOCATION
N_NEAR_HO
OME). Nothinng else happeneed; no one was present. Accep
pt the
incident as iss.
NCVS-CS Pilot
P
Report
I-16
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N2
10-30-09
2)
de in LOCATION_GENER
RAL is “13”-“177” or “36," deteermine if any o
other
If the Precod
NCVS crimee happened in addition
a
to the aattempted theftt. See Section 22, Subsection D
D,
pages 4 throuugh 7 for descrriptions of whatt constitutes ann NCVS crime.
a)
If there is another crim
me other than thhe attempted thheft, delete the aattempted theftt by
changingg screen ATTE
EMPTHEFT to “2." The CA
API instrumentt automatically
deletes the
t entries in sccreens ATTEM
MPTTHEFTW
WHAT through
h
ATTEM
MPTTHEFTIITEMS.
Examplee: An attempt was made to stteal a 9 year oldd household meember's bike fro
om the
parking lot of the mall. (LOCATION
N_GENERAL
L is “15”and Prrecode “29” is
entered in screen LOC
CATION_PAR
RKING.) His 114 year old brother was presen
nt and
attacked
d during the incident. The 14 yyear old reportted the attack an
nd the attempteed
theft of the bike. Keep
p the informatioon about the atttack of the 14 yyear old brotheer.
Delete the
t attempted th
heft of the bikee that belongedd to the 9 year o
old. Enter “2” iin
screen ATTEMPTTH
A
HEFT.
b) If the atttempted theft is
i the only NCV
VS crime that ooccurred, deletee the incident ass
“Out-off-Scope.”
Examplee: An attempt was made to stteal a 9 year oldd household meember’s bike fro
om the
yard of a neighbor's ho
ome. (LOCAT
TION_GENER
RAL is “13” an
nd Precode “199” is
entered in screen LOC
CATION_OTH
HER_HOME
E.) Nothing else was stolen; no
o one
was pressent. Delete th
he incident as ann “Out-of-Scoppe” incident.
b If in
n doubt, enter Referral
R
Code 32 and/or 13 annd refer to HQ..
3
If the precode
p
in screeen ATTEMPT
TTHEFTOWN
NER is “5," takke the followingg steps to correect the
report:
a Reviiew the Summaary and the “Sp
pecify” entry forr this screen.
b Recllassify this entryy to one of the precoded categgories, if possibble.
Exxample: The Suummary states that
t the stolen pproperty actuallly belongs to a nonhouseholdd
meember. Changee item ATTEM
MPTTHEFTO
OWNER from Precode “5” to
o Precode “4." Then,
folllow the instrucctions above forr resolving thesse incidents.
NCVS-CS Pilot
P
Report
I-17
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N2
10-30-09
c
If the “Specify” entry or Summary indicatees that the stoleen property belongs to the
respondent AND
A
a nonhouusehold membeer, take the folloowing steps to correct the report:
1)
If the Summ
mary gives a breaakdown of the pproperty ownedd by the respon
ndent and owneed by
the nonhousehold member,, delete the prooperty that belonnged to the non
nhousehold meember
from screen ATTEMPTTH
HEFTWHAT
T. This informaation is often no
oted in the Sum
mmary.
Change screeen ATTEMPT
TTHEFTOWN
NER to Precodde “1," “2," or “3."
Example: Th
he respondent'ss garage was brroken into. Thee offender(s) atttempted to steaal
several itemss. The propertyy included the rrespondent's tools and a boat b
belonging to a
nonhouseholld member. Deelete the boat fr
from screen AT
TTEMPTHEF
FTWHAT. Ch
hange
screen ATTE
EMPTHEFTO
OWNER to P
Precode (1), “Reespondent onlyy."
2)
If the Summ
mary or the entryy does not give the breakdownn of the properrty owned by th
he
respondent and
a owned by the nonhousehoold member, ennter Referral Co
ode 11 and refeer to
HQ.
Example: Th
he respondent'ss garage was brroken into. Thee offenders atteempted to steall
several itemss belonging to the
t respondent and a nonhoussehold memberr. The propertyy
included toolls and a boat. Do
D not edit thee precodes in sccreen ATTEM
MPTHEFTWH
HAT.
d
4
If in
i doubt, enter Referral Code 11 and refer too HQ.
If in do
oubt, enter Refeerral Code 11 fo
or a nonhousehhold property pproblem, Referrral Code 20 for a
recogniized business property problem
m, and/or Refeerral Code 32 foor a theft or atttempted theft
problem
m and refer to HQ.
H
WHATWASTA
W
AKEN (What Was
W Taken?)
Examine thee Summary and
d Notes and verrify the correct categories werre selected. If n
necessary, changge
screen WHA
ATWASTAKE
EN to agree witth the Notes orr the Summary..
1
If the Summary
S
and Notes
N
do not ind
dicate what wass taken, accept the entries.
2
If the selected Precodee is “11," accep
pt the entry evenn if other itemss besides cash w
were taken.
NCVS-CS Pilot
P
Report
I-18
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N2
3
10-30-09
oubt, enter Refeerral Code 32 an
nd refer to HQ
Q.
If in do
WHOOWNED
W
DSTOLENPROPERTY (Whho Property Belong
ngs To)
Examine thee Summary and
d Notes and verrify the correct categories werre selected. Corrrect as necessaary.
1
If the Summary
S
or No
otes indicate thaat the stolen mooney or properrty belonged to a nonhouseholld
membeer, recognizablee business, scho
ool, etc. (WHO
OOWNEDSTO
OLENPROPE
ERTY is “4”) ddo the
followin
ng:
a Revview the inciden
nt to determine if another NCV
VS crime occurrred. See Sectio
on
2, Subsection D, pages 4 through 7 for descriptioons of what coonstitutes an
NCV
VS crime.
b If no
o other NCVS crime occurred
d, then delete thhe incident as ““Out-of-Scope.””
c If th
here is another NCVS crime in
n conjunction w
with the theft oof the nonhouseehold property,, delete
the theft of the nonhousehold pro
operty from scrreen WHATW
WASTAKEN. IIf necessary, ch
hange
the screen THEFT
T to Precode “2."
“
Ex
xample: The reespondent repo
orts that a friennd's (a nonhoussehold memberr’s) jacket was stolen
fro
om the respond
dent's home durring an illegal entry. The resspondent's pursse was also stolen.
Deelete the theft of
o the jacket fro
om screen WH
HATWASTAK
KEN. Also, ussing informatioon in
thee Notes or Sum
mmary, delete the
t value of thee property (jackket) from screeen
PR
ROPERTYVA
ALUE, if the prroperty value iis known. Keeep the remaininng part of the
inccident that inclludes the illegaal entry of the rrespondent's hoome and the thheft of the
resspondent's pursse.
d If in
n doubt, enter Referral
R
Code 11
1 and refer too HQ.
2
If theree is any indicatio
on the money or
o property beloonged only to a household meember under 122, take
the following steps to correct the rep
port:
a Loo
ok at screens LO
OCATION_G
GENERAL, LO
OCATION_IN
N_HOME andd
LOCATION_NE
EAR_HOME.
1)
NCVS-CS Pilot
P
Report
If the precod
de in LOCATION_GENER
RAL is “11” or ““12," accept th
he incident as is.
I-19
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N2
10-30-09
Example:
2)
A 9 year old household member's bikee was stolen fro
om the respondent's
front yard (LOCATION
N_GENERAL
L is “12” and P
Precode “15” iss
selected in
i screen LOCA
ATION_NEA
AR_HOME.) Nothing else
happened
d; no one was ppresent. Acceppt the incident aas is.
If the Precod
de in LOCATION_GENER
RAL is “13”-“177” or “36," deteermine if any o
other
NCVS crimee happened in addition
a
to the ttheft. See Sectiion 2, subsectio
on D, pages 4
through 7 for descriptions of
o what is an N
NCVS crime.
a)
If th
here is another crime other thhan the theft, deelete the theft b
by changing screeen
TH
HEFT to “2," and
a enter “2” inn screen ATTE
EMPTTHEFT
T. The CAPI
insttrument automaatically deletes tthe entries in sccreens WHATWASTAKEN
thro
ough RECOVE
EREDINSUR
RANCE.
Exaample:
b)
If th
he theft is the only
o
NCVS criime that occurrred, delete the incident as “O
Out-ofSco
ope."
Exaample:
NCVS-CS Pilot
P
Report
A 9 year old hoousehold member's bike was sstolen from thee
parking lot of the mall. (LOC
CATION_GE
ENERAL is “15” and
Precode 29 is eentered in screeen LOCATION_PARKING
G).
His 14 year oldd brother was ppresent and wass attacked durin
ng the
incident. The 14 year old repported the attacck and the theftt of the
bike. Keep thee information aabout the attackk of the 14 yearr old
brother. Delette the theft of tthe bike that beelonged to the 9 year
old. Enter “2”” in screens TH
HEFT and ATTEMPTTHE
EFT.
A 9 year old hhousehold mem
mber's bike wass stolen from thhe
yard of a neigghbor's home. ((LOCATION
N_GENERAL is
“13” and Preccode “19” is enntered in screenn
LOCATION__OTHER_HO
OME.) Nothinng else was stolen;
no one was prresent. Delete the incident ass an “Out-of-Sccope”
incident.
I-20
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N2
10-30-09
b If in
n doubt, enter Referral
R
Code 32 and/or 13 annd refer to HQ..
3
If the Precode
P
in screeen WHOOWN
NEDSTOLEN
NPROPERTY
Y is “5," take thee following step
ps to
correct the report:
a Reviiew the Summaary and the “Sp
pecify” entry forr this screen.
b Recllassify this entryy to one of the precoded categgories, if possibble.
Example: Th
he Summary staates that the stoolen property aactually belongss to a nonhouseehold
member. Ch
hange item WH
HOOWNEDST
TOLENPROP
PERTY from Precode “5” to
o
Precode “4." Then, follow the
t instructionss above for resoolving these inccidents.
c If th
he “Specify” entry or Summaryy indicates thatt the stolen propperty belongs to
the respondent
r
AN
ND a nonhouseehold member, take the follow
wing steps to
corrrect the report:
1)
If the Summ
mary gives a breaakdown of the pproperty ownedd by the respon
ndent AND ow
wned
by the nonho
ousehold memb
ber, delete the pproperty and thhe value of that property that
belonged to the
t nonhouseh
hold member froom screen WH
HATWASTAKE
KEN and screen
n
PROPERTY
YVALUE, if possible.
p
This innformation is ooften noted in tthe Summary.
Change screeen WHOOWN
NEDSTOLEN
NPROPERTY
Y to Precode “1,"”2," or “3."
Example: The
T respondent’s garage was bbroken into. S
Several items w
were stolen. Thhe
stolen propeerty included th
he respondent’ss tools valued aat $100 and a bboat belongingg to a
nonhouseholld member. Th
he value of thee boat was $2500. Delete the bboat from screeen
WHATWASTAKEN. Ch
hange screen W
WHOOWNED
DSTOLENPR
ROPERTY to
Precode (1), “Respondent only."
o
Changee the value in sscreen PROPE
ERTYVALUE
E to
$100.
2)
If the Summ
mary or the entryy does not give the breakdownn of the properrty owned by th
he
respondent and
a owned by the nonhousehoold member, ennter Referral Co
ode 11 and refeer to
HQ.
Example: Th
he respondent'ss garage was brroken into. Sevveral items weree stolen. The sttolen
property inclluded the respo
ondent's tools annd a boat belonnging to a nonh
household mem
mber.
Delete the bo
oat from screen
n WHATWAST
TAKEN. Channge screen
WHOOWN
NEDSTOLEN
NPROPERTY to Precode (1), “Respondent only." Do nott edit
the value in screen
s
PROPE
ERTYVALUE .
NCVS-CS Pilot
P
Report
I-21
endix I
Appe
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
Census Instructio
PROJECT 7523008
7523008-9009-C
SECTION
N2
10-30-09
n doubt, enter Referral
R
Code 11 and refer to H
HQ.
d If in
5
If in do
oubt, enter Refeerral Code 11 fo
or a nonhousehhold property pproblem, Referrral Code 20 for a
recogniized business property problem
m, and/or Refeerral Code 32 foor a theft or atttempted theft
problem
m and refer to HQ.
H
PROPERTYVA
ALUE (Value of
o Stolen Propeerty)
The FR sho
ould have entereed the value of the stolen propperty, excludingg any stolen cassh, checks, or crredit
cards. Noting this, edit as follows:
1
If theree is any indicatio
on that the amo
ount entered is only the value of the stolen ch
hecks or credit cards,
delete the
t entry.
2
If theree is any indicatio
on that the amo
ount entered is the value of thhe stolen properrty plus the facee
value of checks or cred
dit cards:
a Chaange the entry to
o the value of the
t property onnly.
b If th
he value of the property
p
canno
ot be determineed, delete the enntry in screen
PR
ROPERTYVA
ALUE.
3
If theree is any indicatio
on that the amo
ount entered inncludes the valuue of the nonho
ousehold property,
examin
ne the Summaryy and Notes to see if you can ddetermine whatt that value is an
nd subtract that
amountt from the amo
ount entered. Iff unable to deteermine, enter R
Referral Code 111 to refer to HQ
Q.
4
If the screen PROPE
ERTYVALUE is blank, or a bblind “Don’t Knnow” or a blindd “Refused” waas
entered
d, and an entry should
s
have been made, exam
mine the Summaary and any oth
her Notes to tryy to
mine,
determiine what the en
ntry should be. If able to deterrmine, enter thhat amount. If uunable to determ
enter Referral
R
Code 244 to refer to HQ
Q.
RECOVERED
R
DCASHVALUE
E (Value of Reccovered Properrty)
If there is an
ny indication th
hat the amount entered does nnot include the vvalue of the reccovered properrty,
edit as follow
ws:
1
Changee the entry to in
nclude the valuee of the propertty recovered.
2
If the value
v
of the reco
overed propertyy cannot be dettermined, delette the entry in screen
RECO
OVEREDCASH
HVALUE.
NCVS-CS Pilot
P
Report
I-22
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N2
10-30-09
REASONREPO
R
ORT (Reason Incident
I
Reporrted To The Poolice)
If Precode (11),
(
“Stop or prevent
p
THIS in
ncident from haappening” is seelected and the respondent was not
present (HH
HMEMBERP
PRESENT = “2”)
“ follow the instructions beelow.
1
If only Precode “11” is
i selected, deseelect Precode “ 11” and then seelect Precode ““22."
2
If Preco
ode “11” is seleected in conjunction with otheer precodes, desselect Precode “11." If Precode
“11” is also selected in
n the screen RE
EPORTIMPO
ORTANT, deleete the entry in
REPO
ORTIMPORTA
ANT.
SE
ERIESNUMT
TIMES (Numb
ber Of Times In
n This Series)
If this screen
n appears and the
t incident hass not already beeen referred as a “series of inccidents," enter
Referral Cod
de 28 and referr to HQ.
F. VERIF
FICATION
Verify
V
the editingg/coding operaation on a 100%
% basis. Follow
w the instructionns for reviewin
ng incidents as
sp
pecified in Sectiion 2, Subsectio
ons A through E.
1
Verifiers aree not allowed to
o verify their ow
wn work.
2
Verify as specified in the cllerical E/C insttructions.
3
When enteriing a case, durin
ng the Verificattion Stage, pleaase note any refferral reasons allready entered ffor an
incident. Iff an incident hass already been referred
r
there w
will be a referral
al code (2 – 40) in the column
EF CODE” in the
t screen INC
CROSTER_EC
C.
labeled “RE
a
If durin
ng your review of the incident you agree withh the referral reaason, you do no
ot need to re-en
nter
that refferral reason.
b
If durin
ng your review of the incident you determinee another reasonn to refer the in
ncident, follow the
procedu
ures for referrin
ng incidents to HQ, as specifieed in Section 1,, Subsection I.
c
If durin
ng your review of the incident you disagree w
with the referrall reason, complete your review
w of all
inciden
nts in the case. In the Notes seection of WebC
CATI discuss anny discrepanciees between the first
round review
r
and the verification revview.
NCVS-CS Pilot
P
Report
I-23
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N2
G.
10-30-09
DISPOSITION
D
1
The NP
PC is to notify Kathryn
K
Chezaa, DSD/CSB, att 301 - 763 - 39925 (or via e-maail), when editin
ng and
coding is completed th
hrough verification.
2
If there are any questions regarding
r
this procedure,
p
conttact Kathryn Chheza.
3
If the month
hly closeout can
nnot be met, no
otify Kathryn C
Cheza as soon aas possible.
4
If problems are encouuntered accessin
ng cases (i.e. casses are locked oor are already in
n use), notify K
Kathryn
Cheza as
a soon as posssible.
NCVS-CS Pilot
P
Report
I-24
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
A
ATTACHMEN
NT A-1
LIST OF
O CRIME RE
ECLASSIFICA
ATION CODES
S AND
LIST OF REF
FERRAL REA
ASON CODES
Crime Re
eclassification
n Codes for the
e Catchall Que
estions
11
12
13
14
15
16
17
18
19
20
21
22
23
Rape
mpted Rape
Attem
Robb
bery
Attem
mpted Robbery/Threatened Robbery
R
Assau
ult
Attem
mpted Assault/ Threatened Asssault
Burgllary
Attem
mpted Burglary
Larceeny (Household
d or Personal)
Attem
mpted Larceny (Household or
Perso
onal)
Moto
or Vehicle Thefft
Attem
mpted Motor Vehicle
V
Theft
Moto
or Vehicle Accid
dent
NCVS-CS Pilot
P
Report
24
25
26
27
28
29
30
31
32
33
I-25
Vaandalism (Againnst Household or
Hoousehold Membber’s Property)
Proowlers/Peepingg Toms
Cri
rime Against Hoousehold, Otheer than
Abbove
Noot a Crime
Cri
rime Against Soomeone Else orr Society
Unnable to Classify
fy
Unnwanted Sexuall Contact Again
nst a
Hoousehold Membber (with & witthout
foorce)
Haate Crime Againnst Household
Cyybercrime Againnst Household
Ideentity Theft Aggainst Househo
old
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
A
ATTACHMEN
NT A-2
List of Refferral Reasonss
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
used Screeners
Refu
Wheere Happened
Misssing Incident
Hou
usehold Respon
ndent Under 188
Unaacceptable Proxxy
Pressence
Rape/Sexual Assau
ult
Nam
me Entered in
HHMEMHARME
ED_NAMES
Catcch-All Codes
Weaapon
Non
nHousehold Pro
operty/Joint Prroperty
Vandalism
Child in HH Underr 12/Bike or To
oy
Stolen
Whaat Actually Hap
ppened
Threeat/Try Attack
Policce Officer Offeender
NCVS-CS Pilot
P
Report
17. Fraaud
18. Inccident Out-of-SScope
19. Duuplicates
20. Coommercial Crim
me/Recognizablle Business
21. NC
CVS Crime
22. Tellephone Threatts
23. Peeeping Tom/Exxhibitionist/Arsson
24. Mi scellaneous (Kiidnapping/Hit & Run)
25. Shoould Not Havee Been Referredd
26. Enntrusted/Borrow
wed Property
27. Moore than One Sccreener Refused – Type A
or Type Z
28. Serries of Incidentts
30. Perrsonal Crime w
w/Contact Repo
orted By
Perrson Other thaan Victim
31. Atttempted Attackk & Attack Exccept Boxes
22 and 24
32. Thheft & Attempteed Theft Probleems
39. Racce Equals Otheer - Specify
40. Deelete Incident
I-26
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
A
ATTACHMEN
NT B-1
IN
NSTRUCTION
NS FOR EDITIING CONTRO
OL CARD SCR
REEN RACE
A
INTRO
ODUCTION
The prrecoded race caategories for scrreen RACE aree: (1), “White”;; (2), “Black, orr African Amerrican”; (3), “Am
merican
Indian
n or Alaska Natiive”; (4), “Asian
n”; (5) “Native Hawaiian or O
Other Pacific Isllander”; and (6)), “Other-speciify.”
In add
dition to allowin
ng multiple-racee reporting, thee race question aalso allows for write-in entriess to describe th
he race
when Precode
P
(6), “O
Other-specify” is
i selected.
At the request of our sponsor, the Bureau
B
of Justicce Statistics (BJSS), all write-in eentries of race m
must be revieweed,
when “Other-specify
“
,” Precode (6) is
i selected, in order to determiine if the “otheer race” can be classified into o
one of
the fivve precoded race categories. To
T accommodatte this request, the E/C instruument will displlay the screen
RACE
EROSTER_EC
C when Precod
de (6) “Other-sp
pecify” is selectted in the RAC
CE screen. Thee
RACE
EROSTER_EC
C displays the line
l number un
nder the header (“LNO”), resppondent’s namee, relationship to
o the
referen
nce person (“RE
EL”), Hispanicc origin (“ORG
G”), the selectedd race precode ((“RACE”), andd the other speccified
race (““RACE SPECIF
FY”). It is this information yo
ou will review aand edit as desccribed in Section C of this
attachm
ment. This screen allows the NPC
N coder to reclassify
r
the raace entry speciffied in the “RAC
CE-SPECIFY””
column, if necessary, to one of the five
f precoded raace categories.
B
SITUA
ATIONS FOR WHICH THE
E RACEROST
TER_EC SCRE
EEN WILL AP
PPEAR IN THE
E E/C
INSTR
RUMENT
This sccreen will appear only when one or more resp
pondents on thhe household rooster selected P
Precode 6 in thee
screen
n RACE, regard
dless if they rep
ported incidentss. Only responddents on the hoousehold rosterr who reported a race
other than
t
or in addittion to one of the
t five precodeed race categoriies can be editeed at the RACE
EROSTER_EC
C
screen
n.
C
REVIE
EWING AND
D EDITING TH
HE OTHER RACE
RA
WRITE-IIN ENTRIES
1
When
W
screen RA
ACEROSTER
R_EC appears, review
r
the “Othher-specify” wrrite-in entry(ies)) listed in the
“R
RACE SPECIF
FY” column forr each person.
2
Use
U the situation
nal guidelines in
n the table on th
he next two pagges to determinne how to edit tthe “Other-speccify”
raace for each respondent, when
n necessary. If in
i doubt, refer the case to HQ
Q.
NCVS-CS Pilot
P
Report
I-27
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
A
ATTACHMEN
NT B-2
3 If you are
a able to reco
ode all of the write-in
w
race en
ntries, enter thee appropriate Prrecode(s) 1-5 aand delete Preccode
6. If the
t appropriatee Precode 1-5 is already seleccted, you only nneed to remove Precode 6.
4
a
When able to recode all of the write-in entries:
e
If “Hisspanic” and “Inndian (Non-Naative Americann)” are the
write-in racce entries and Precode
P
(6) is the
t only precodde selected, desselect Precode (6) and select Precode
(1) “White”” for “Hispanicc” and Precode (4) “Asian” foor “Non-Nativee American.”
b
When able to recode all of the write-in entries
e
and thosse categories aare already seleected: If Precoodes (1)
and (4) are selected and Precode (6) is also marked witth write-in entrries of “Hispannic” and “Indiaan (NonNative Ameerican),” keep Precodes (1) an
nd (4) and des elect Precode ((6).
Iff only part of th
he write-in entryy can be recodeed, recode whatt you can to thee appropriate precode and leavve
Prrecode (6) filled
d. For examplee, if “Hispanic” and “Americann” are the writee-in entries, reccode “Hispanic”” to
Prrecode (1) and leave the “Other-specify” Precode (6) as is.
5
6
If you aree in doubt or caannot recode th
he “Other-speccify” entry refeer to HQ using Referral Code 39.
After
A
editing the respondent’s race
r codes, wheen applicable, thhe instrument w
will progress to any catchall
qu
uestions. Please refer to Sectio
on 1, Subsectio
on E5, for the oorder in which tthe instrument progresses thro
ough
an
ny additional “O
Other-specify” race entries and
d “Catchall” quuestions for a household.
SITUA
ATION
1
2
3
4
IF WRITE-IN
W
ENTRY
Y(IES) IS/ARE
E:
“R
Refused," “R," “Ref,"
“
“Blind Refusal,"
R
(w
which is displayeed as a blue exclamation
po
oint)
OR
R
“D
Don’t Know, “D
D," “DK”
“N
None of your buusiness” OR “N
None of
thee above”
“H
Human Race," “Bi-racial,"
“
“Duual,"
“M
Mixed”
“ Cuban,"
C
“ Dom
minican,"
“ Hispanic,"
H
“ Latino,"
“ Mexican,"
M
“Pueerto Rican," OR
R
“ Spanish”
S
NCVS-CS Pilot
P
Report
I-28
THEN:
D
Delete Precode (6) and then co
ontinue to edit.
D
Delete Precode (6) and then co
ontinue to edit.
D
Delete Precode (6) and then co
ontinue to edit.
R
Recode to Precoode (1), “Whitee” and deselect
P
Precode (6).
endix I
Appe
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
A
ATTACHMEN
NT B-3
IF WRITE-IN
W
ENTRY
Y(IES) IS/ARE
E:
SITUA
ATION
THEN:
5
Blaank (No write-iin entry)
R
Refer to HQ; R
Referral Code 399.
6
Sin
ngle or multiplee entry(ies) are not
n listed
in situations 1 - 5 of this table.
(SSINGLE OR M
MULTIPLE EN
NTRY(IES))
R
Review the “Raace Recode List” in this
aattachment. If aall of the write--in entries (singgle
eentry or multiplle entries) are listed on the “Raace
R
Recode List," deelete Precode (6). Then selectt the
ccorresponding R
Race category(ies), Precode (1))
thhrough (5), from the “recode”” column on th
he
““Race Recode L
List” table in Seection D and
ccontinue to editt.
IIf none of the w
write-in entries (single entry orr
m
multiple entries)) are listed on tthe “Race Reco
ode
L
List," refer to H
HQ and then co
ontinue to edit.
(M
MULTIPLE E
ENTRIES)
R
Review the “Raace Recode List” in this
aattachment. If oonly some of th
he entries are listed
oon the “Race Reecode List," reccode those writte-in
eentries by selectting the approp
priate race
ccategory(ies), Prrecode (1) throuugh (5) from th
he
““recode” colum
mn on the “Racee Recode List”
taable in Section D. Leave Preccode (6) as a
sselected race.
SSince there are ssome remainingg write-in entriees
thhat cannot be rrecoded, refer tthese cases to H
HQ
aand continue too edit.
7
Basic racial and ethnic
e
categoriees are defined as
a follows:
a
“White” - A person havingg origins in anyy of the originall peoples of Euurope, North Affrica, or the Miiddle
East.
b “Black” - A person
p
having origins
o
in any of
o the black raccial groups of A
Africa.
NCVS-CS Pilot
P
Report
I-29
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
c
A
ATTACHMEN
NT B-4
“American Indian or Alaaskan Native” - A person havving origins in aany of the original peoples of N
North
America, an
nd who maintain
ns cultural iden
ntification throuugh tribal affiliaations or comm
munity recognition.
d “Asian or Pacific Islander”” - A person haaving origins inn any of the orriginal peoples of the Far East,
Southeast Asia,
A
Indian sub
bcontinent, or the
t Pacific Islaands. This areaa includes, for example, Chinna,
India, Japan
n, Korea, the Philippine Islan
nds, and Samoaa.
e
D
“Hispanic”” - A person off Mexican, Puerrto Rican, Cubaan, Central or South American
n or other Spaniish
culture or orrigin, regardlesss of race. For the
t NCVS, “Hiispanic” shouldd be coded as “W
White”.
RACE
E CODE LIST
Use th
he chart below to
t recode cases falling under situation 6 in thhe table on the ppreceding page,, regardless of rrace
entriess already selecteed. Unless a racce category is already selected (Precodes 1–5)), key the preco
ode for the racee that
corresp
ponds to the write-in
w
entry usiing the chart beelow. For exam
mple, if the writte-in entry is “A
Aleut," select Prrecode
(3) “American Indian
n/Alaska Nativee.” Dependingg on whether thhere are other w
write-in entries yyou may also bee
requireed to delete Preecode (6). Refeer to the table on
o the precedingg pages for speecific situations and edit procedures
for theese situations.
NOTE
E:
The “R
Race Recode Lisst” is listed alph
habetically by thhe write-in entrry.
NCVS-CS Pilot
P
Report
I-30
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
A
ATTACHMEN
NT B-5
Race
e Recode List
If the “O
Other-specify”
” write-in entryy is...
Then recoode to race caategory:
African Nation,
N
Ethnic Group, or Trib
be
2 (Black oor African Amerrican)
African American
A
2 (Black oor African Amerrican)
Afro Am
merican
2 (Black oor African Amerrican)
Alaska Native
N
3 (Americcan Indian/Alasska Native)
Aleut
3 (Americcan Indian/Alasska Native)
American
n Indian
3 (Americcan Indian/Alasska Native)
Anglo-Saaxon
1 (White)
Arab
1 (White)
Asian Ind
dian
4 (Asian)
Asian
4 (Asian)
Black
2 (Black oor African Amerrican)
Cambodiian
4 (Asian)
Caucasian
n
1 (White)
Chamorrro
5 (Native Hawaiian/Othher Pacific Islan
nder)
Chinese
4 (Asian)
Eskimo
3 (Americcan Indian/Alasska Native)
European
n
1 (White)
Filipino
4 (Asian)
German
1 (White)
Greek
1 (White)
Guamaniian
5 (Native Hawaiian/Othher Pacific Islan
nder)
Haitian
2 (Black oor African Amerrican)
Indian (n
non-Native American)
4 (Asian)
Italian
1 (White)
Jamaican
n
2 (Black oor African Amerrican)
Japanese
4 (Asian)
NCVS-CS Pilot
P
Report
I-31
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
A
ATTACHMEN
NT B-6
Race
R
Recode List
L (Continue d)
If the “Oth
her-specify” write-in
w
entry is...
i
Then recodee to race categ
gory:
Kenyan
2 (Black or A
African Americaan)
Korean
4 (Asian)
Lebanese
1 (White)
Malaysian
4 (Asian)
Native Haw
waiian
5 (Native Haawaiian/Other P
Pacific Islanderr)
Native Am
merican
3 (American Indian/Alaska Native)
Near-Easteerner
1 (White)
Negro
2 (Black or A
African Americaan)
Nigerian
2 (Black or A
African Americaan)
Other Pacific Islander
5 (Native Haawaiian/Other P
Pacific Islanderr)
Pakistani
4 (Asian)
Polish
1 (White)
Samoan
5 (Native Haawaiian/Other P
Pacific Islanderr)
Scotch-Irish
1 (White)
Vietnamesee
4 (Asian)
West Indian
2 (Black or A
African Americaan)
White
1 (White)
NCVS-CS Pilot
P
Report
I-32
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
A
ATTACHMEN
NT C-1
INST
TRUCTIONS FOR
F
CRIME IN
NCIDENT RE
EPORT SCREE
EN WEAPON
N
If Precode 4, 5, or 6 is seleected in the scrreen WEAPON
N, edit as outlinned in this attacchment.
A
GENERAL
G
INSSTRUCTIONSS
1 Examine the Summary
S
and any
a Notes mad
de by the FR too help determinne the appropriiate entries.
2 If the weapon
n cannot be deteermined, but Precode 4 and/oor 5 is selectedd, accept the weeapon.
3 If the weapon
n is any object not
n listed in Weapon
W
Types 1 through 3 in S
Section B of thhis attachment, enter
Referral Co
ode 10 and refeer the incident to
t HQ.
4 If you have an
ny doubts abou
ut an entry, also
o use Referral Code 10 and reefer the incidennt to HQ.
NCVS-CS Pilot
P
Report
I-33
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
B
A
ATTACHMEN
NT C-2
WEAPON
W
ENT
TRY TYPES
Using
U
the charts below, determine the Weapon
n Entry Type (11-3), then edit sscreens WEAP
PONPRESEN
NT
th
hrough MEDIC
CALEXPENS
SES according to
t the instructioons provided foor that type of w
weapon entry. If in
do
oubt, refer the incident
i
to HQ
Q.
1 TYPE 1 -
aciid
afrro combs
ash
h trays
If the weapon is any object listed below, coonsider the object to be a Typ
pe 1 entry and ffollow
the instructio
ons for Type 1 edits in Sectionn C.
1/
caake cutter
ch
hains
ch
hair
kknife
lyye
nnight stick
nnunchucks (marrtial
aarts weapon,
cconsisting of a
cchain between ttwo
s ticks)
1/
1/
1/
shovel
stick
straight razor
axe
cllub
1/
baaseball bat
crrowbar
ppipe
1/
beelt
daarts
ppitchfork
w
wrench
ppliers
U
Unknown (weaapon
p
present, but vicctim
n
not sure what iit
w
was)
1/
billly club
1/
1/
1/
blaackjack
bo
oard
bo
ombs
bo
ottle
brass knuckles
briick
broom
exxplosives
1/
gllass
guun
haammer
haatchet
ice pick
jack handle
kaarate rope
1/
1/
1/
1/
ttire iron
1/
ttree limb
ppoison
ppool stick
raake
roock
s cissors
s crew driver
s hoe (held in haand)
1/ If there
t
is an indicattion that the objectt was thrown, folloow the instructionss provided for Typpe 2 edits in Sectiion C.
NCVS-CS Pilot
P
Report
I-34
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
2 TYPE 2 -
A
ATTACHMEN
NT C-3
If the entry is any object listted below and tthere is an indiccation that it w
was thrown, follo
ow the
instructions provided
p
for Tyype 2 edits in SSection C.
baseball
b
bat
belt
b
board
b
bottle
b
brick
b
broom
b
3 TYPE 3 -
rock
screw drriver
shovel
stick
tree limb
b
unknown
n
If the entry is any object listted below, do nnot consider thee object to be a weapon and fo
follow
the instructio
ons provided fo
or Type 3 edits in Section C.
animals
a
(dog, caat, etc.)
BB
B guns (unlesss used as a club
b)
Chloroform
C
food
f
NCVS-CS Pilot
P
Report
chair
darts
glass
pliers
pool stickk
rake
mace
part of thhe body (hands,,
feet, etc.)
small emppty can/bottle (no
injury)
tear gas
I-35
tear gass gun
water b
balloons
taser/ sstun gun
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
C
A
ATTACHMEN
NT C-4
DETAILED
D
IN
NSTRUCTIONSS
1
TYPE 1 ED
DITS
a
b
If the victim
v
was actuaally attacked (A
ATTACK = 1, oor HOWATTA
ACK = 11-14, 16-17, 19-20, o
or 2224, or INJURY
I
= 12--21), edit WEA
APONPRESEN
NT through M
MEDICALEXP
PENSES as fo
ollows:
1)
WE
EAPONPRES
SENT - Verifyy that Precode 1 is entered, or correct as neceessary.
2)
WE
EAPON - Verrify that Precod
de 4, 5, or 6 is enntered, or correect as necessaryy.
3)
AT
TTACK - Veriffy that Precode 1 is entered, orr correct as neccessary.
4)
HO
OWATTACK through MED
DICALEXPEN
NSES - Acceptt all entries.
If the victim
v
was not actually
a
attacked
d, but an attem
mpt was made (T
TRYATTACK
K=1,
HOWT
TRYATTACK
K = 11-24, or HOWATTACK
H
K = 15, 18 or 211 only), edit as follows:
1)
WE
EAPONPRES
SENT - Verifyy that Precode 1 is entered, or correct as neceessary.
2)
WE
EAPON - Verrify that Precod
de 4, 5, or 6 is enntered, or correect as necessaryy.
3)
AT
TTACK - Veriffy that Precode 2 is entered, orr correct as neccessary.
4)
TR
RYATTACK - Verify that 1 iss entered, or coorrect as necessaary.
5)
HO
OWTRYATTA
ACK - Verify th
hat appropriatee codes are enteered, or correctt as necessary. If
theere are entries in
n HOWATTA
ACK then code HOWTRYAT
TTACK using the following
con
nversion:
a)
If HOWATT
TACK = 15, seelect Precode 1 8 in HOWTRY
RYATTACK.
b) If HOWATT
TACK = 18, seelect Precode 1 9 in HOWTRY
RYATTACK.
c)
NCVS-CS Pilot
P
Report
If HOWATT
TACK = 21, seelect Precode 220 in HOWTRY
RYATTACK.
I-36
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
c
A
ATTACHMEN
NT C-5
If the victim
v
was not actually
a
attacked
d but was threaatened (THRE
EATEN=1, HO
OWTHREAT
TEN =
11-17 or
o 22, or HOW
WATTACK indiicates that the vvictim was not actually attackeed, but an attem
mpt
was maade), edit as folllows:
1)
WE
EAPONPRES
SENT - Verifyy that Precode 1 is entered, or correct as neceessary.
2)
WE
EAPON - Verrify that Precod
de 4, 5, or 6 is enntered, or correect as necessaryy.
3)
AT
TTACK througgh TRYATTACK - Verify thaat Precode 2 is entered, or corrrect as necessaary.
4)
TH
HREATEN - Verify
V
that Precode 1 is entereed, or correct aas necessary.
5)
HO
OWTHREAT
TEN - Verify th
hat Precode 11--17 or 22 is enttered, or correctt as necessary.
d If the object was used only
o
to gain acccess into a buillding or motor vehicle, and no personal
confron
ntation occurreed between the offender and tthe victim, editt as follows:
2
1)
WE
EAPONPRES
SENT - Select Precode 2.
2)
AT
TTACK througgh THREATE
EN - Verify thaat Precode 2 is eentered, or corrrect as necessarry.
3)
WH
HATHAPPEN
N - Verify that the appropriatte codes are enttered, or correcct as necessary.
TYPE 2 ED
DITS
a If the object was thrown
n at the victim..
1)
Co
onsider the objeect a weapon an
nd the victim atttacked if:
a)
NCVS-CS Pilot
P
Report
The victim was
w injured (INJJURY = 12-211) AND the offfender is at least 12 years of agge, or
the age of the offender is un
nknown (SING
GOFFENDER
RAGE,
MULTOFF
FENDERYOU
UNG, MULTO
OFFENDERO
OLD = 2-7)
I-37
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
A
ATTACHMEN
NT C-6
b) The victim was
w seriously injjured (INJURY
Y = 15-20) andd the offender iss any age.
NOTE:
If the offendeer is under 12 yyears of age andd the victim wass not seriously
injured, see in
nstructions in p aragraph C2a(22)(b), Attachmeent C-6.
Edit WEAPO
ON through MEDICALEXP
M
PENSES as foollows:
2)
1.
WEAP
PON - Verify th
hat Precode 4, 55, or 6 is entereed, or correct ass necessary.
2.
ATTACK - Verify thaat Precode 1 is entered, or corrrect as necessaary.
3.
HOWA
ATTACK thro
ough MEDICA
ALEXPENSESS - Accept all eentries.
Do
o not consider the
t object to bee a weapon, butt consider an atttempted attackk on the victim or the
vicctim threatened
d with harm if th
he offender is aat least 12 yearss of age and thee victim is not
injured.
NOTE:
If age of offeender cannot bee determined, cconsider the offfender to be att least
12 years of ag
ge.
a)
If the victim was not attacked in some othher way (HOWA
ATTACK = 200 only, or 24
indicates a th
hrown object, only),
o
edit screenns as follows:
1.
WEAP
PONPRESEN
NT - Select Preccode 2.
2.
ATTACK - Verify thaat 2 is entered, or correct as neecessary.
3.
TRYAT
TTACK or TH
HREATEN - V
Verify that 1 is entered, or corrrect as necessaary.
4.
HOWT
TRYATTACK
K or HOWTHR
REATEN - V
Verify that 21 is entered, or corrrect
and acccept any entriess in 11-16, 21-244. Delete Precodes 17, 18, 199, and/or 20 if
entered
d.
b) If the victim was attacked in
n some other w
way (Precode 200 and/or any otther precodes aare
entered in HOWATTACK
H
K), edit screen W
WEAPONPRE
ESENT througgh
MEDICALE
EXPENSES as
a follows:
NCVS-CS Pilot
P
Report
I-38
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
b
A
ATTACHMEN
NT C-7
1.
WEAP
PONPRESEN
NT - Select Preccode 2.
2.
ATTACK - Verify thaat 1 is entered, or correct as neecessary.
3.
HOWA
ATTACK thro
ough MEDICA
ALEXPENSESS - Accept all eentries.
If the object
o
was throw
wn at a house or
o car or if the ooffender is undder 12 years of aage and the victtim
was nott seriously injurred:
1)
Fo
ollow the instructions in 2a (Atttachment C-5) for objects thrrown at the victtim if there is
eviidence that a seerious attempt was
w made to atttack or injure a person in the h
house or car. F
For
exaample:
a)
Respondent's ex-husband throws
t
bricks tthrough the winndow by whichh respondent iss
standing.
b) The car in wh
hich responden
nt is riding is reepeatedly bombbarded with rocks and bottles dduring
a street riot.
c)
2)
The offenderr is deliberately
y trying to injuure the victim oor cause him too lose control oof his
car by throw
wing object(s) at
a the car.
If the victim was not attacked orr threatened in some other waay, including atttempted attackss, edit
as follows:
a)
WEAPONP
PRESENT - Select Precode 22.
b) ATTACK th
hrough THREA
ATEN - Verify
fy that 2 is enterred, or correct as necessary.
c)
3)
WHATHAP
PPEN - Accep
pt all entries.
If the victim was attacked in som
me other way, eedit screens WE
EAPONPRES
SENT through
ME
EDICALEXP
PENSES as folllows:
a)
WEAPONS
SPRESENT - Select Precode 2.
b) ATTACK - Verify
V
that 1 is entered, or corrrect as necessaary.
c)
NCVS-CS Pilot
P
Report
HOWATTA
ACK through MEDICALEX
M
XPENSES - Acccept all entries.
I-39
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
4)
A
ATTACHMEN
NT C-8
If the victim was not attacked, but
b was threatenned in some othher way, edit ass follows:
a)
WEAPONP
PRESENT - Select Precode 22.
b) ATTACK th
hrough TRYAT
TTACK - Veriffy that 2 is enteered, or correct as necessary.
c)
THREATE
EN - Verify thatt 1 is entered, oor correct as necessary.
d) HOWTHRE
EATEN - Acccept any entriess in 11-16, 21-244. Delete entriies of 17, 18, 199
and/or 20.
3
TYPE 3 ED
DITS
a
b
If the victim
v
was actuaally attacked (A
ATTACK = 1, H
HOWATTAC
CK = 11-24, or INJURY = 122-21),
edit scrreens WEAPON
NPRESENT through MED
DICALEXPEN
NSES as follow
ws:
1)
WE
EAPONPRES
SENT - Select Precode 2.
2)
AT
TTACK - Veriffy that 1 is enteered, or correct as necessary.
3)
HO
OWATTACK through MED
DICALEXPEN
NSES - Acceptt all entries.
If the victim
v
was not actually
a
attacked
d but an attemppt was made (T
TRYATTACK=
=1,
HOWT
TRYATTACK
K = 11-23, or HOWATTACK
H
K = 15, 18 or 211), edit as follow
ws:
1)
WE
EAPONPRES
SENT - Select Precode 2.
2)
AT
TTACK - Veriffy that 2 is enteered, or correct as necessary.
3)
TR
RYATTACK - Verify that 1 iss entered, or coorrect as necessaary.
4)
HO
OWTRYATTA
ACK - Verify th
hat appropriatee codes are enteered, or correctt as necessary. If
theere are entries in
n HOWATTA
ACK, then codee HOWTRYAT
ATTACK using the following
con
nversion:
a)
NCVS-CS Pilot
P
Report
If HOWATT
TACK = 15, en
nter 18 in HOW
WTRYATTAC
CK.
I-40
Appe
endix I
Census Instructio
ons for Using The Automateed NCVS Editing And Coding
g Blaise Instru
ument
PROJECT 7523008
7523008-9009-C
SECTION
N1
A
ATTACHMEN
NT C-9
b) If HOWATT
TACK = 18, en
nter 19 in HOW
WTRYATTAC
CK.
c)
c
d
If HOWATT
TACK = 21, en
nter 20 in HOW
WTRYATTAC
CK.
If the victim
v
was not actually
a
attacked
d, but was threaatened (THRE
EATEN= 1, H
HOWTHREAT
TEN
= 11-244, or HOWAT
TTACK indicatees that the victiim was not actuually attacked, b
but an attempt w
was
made), edit as follows::
1)
WE
EAPONPRES
SENT - Select Precode 2.
2)
AT
TTACK througgh TRYATTACK - Verify thaat 2 is entered, or correct as neecessary.
3)
TH
HREATEN - Verify
V
that 1 is entered, or corrrect as necessaary.
4)
HO
OWTHREAT
TEN - Accept entries
e
of 11-166, 21-24. Delete entries of 17, 18, 19 and/or 20.
If the victim
v
was not attacked
a
and no
ot threatened, eedit as follows:
1)
WE
EAPONPRES
SENT - Select Precode 2.
2)
AT
TTACK througgh THREATE
EN - Verify thaat 2 is entered, oor correct as neecessary.
3)
WH
HATHAPPEN
N - Accept all entries.
e
NCVS-CS Pilot
P
Report
I-41
Appendix I
How to Use the Weights
Appe
endix I
H
How to Use Weights
A few exaamples of ho
ow to use thee weights are given below
w. Rates and oother estimattes should bee
calculated
d separately for
f Approach
hes 1 and 2, since
s
the sam
mple for eachh approach w
weights up to the
same pop
pulation.
n
Example 1.
perty crime victimization
v
n rate =
Prop
HHHPSTWT
i 1
* Xi
i
x 1000
n
HHPSTWTT
i 1
i
where Xi = number of
o property crrime victimizzations reporrted for the i--th householld, HHPSTW
WTi is
h
weight
w
for thee telephone screener, and n is the num
mber of households
the final household
respondin
ng to the houusehold telep
phone screen
ner (HHTSCR
RRESP=1). T
The set of reeplicate weigh
hts to
use for vaariance estim
mation is HHR
REPPSTWT
T1-HHREPP
PSTWT80.
n
Example 2.
Viollent crime vicctimization rate
r =
FINNPERSWT
i 1
i
* Zi
x 1000
n
FFINPERSWTT
i 1
i
where Zi = number of violent crim
me victimizattions reporteed by the i-thh person, FIN
NPERSWTi iis the
final persson weight fo
or the i-th person, and n is
i the numbeer of persons responding to the intervview
(PERSRE
ESPSTATUSS=1). The seet of replicatee weights to uuse for variaance estimatio
on is
FINPER
RSWT1-FINP
PERSWT80.
n
Example 3.
Total num
mber of victim
mizations =
MWT
VICTIM
j
j 1
i the victimiization weigh
ht for the j-thh incident connsidered to b
be inscope fo
or the
Where VIICTIMWTj is
NCVS (IN
NSCOPE=11). The set off replicate weeights to use for variance estimation is VICTIMW
WT1VICTIMW
WT80.
Example 4.
n
Perceent of violent incidents reeported to th
he police =
INCIDENNTWT
j 1
I-1
* REP
PORT j
x 100
n
INCCIDENTWT
j1
NCVS-CS Pilot
P
Report
j
j
Appe
endix I
H
How to Use Weights
NCIDENTW
WTj is the incident weightt for the j-th rreported inciident, n is th
he total numb
ber of
Where IN
violent, in
nscope incideents reported
d, and REPO
ORTj is an inddicator (creatted by the usser) of wheth
her
the j-th crrime was rep
ported to the police based
d on Q115 frrom the interrview. The seet of replicatee
weights to
o use for varriance estimation is INCIDENTWT1 – INCIDEN
NTWT80.
A list of the
t weights and
a their corrresponding replicate
r
weigghts is given below:
HHBASE
EWT = Household
H
baase weight fo
or Approach 1 and for Appproach 2 beefore subsam
mpling
nonrespondin
n
ng household
ds for the maail screener (rreplicate weigghts
HHREPBWT
H
T1-HHREPB
BWT80).
HHBASE
EWT2 = Household
H
baase weight fo
or Approach 2 after subsaampling for n
nonrespondiing
households
h
fo
or the mail sccreener (repliicate weightss HHREPBW
WT2_1HHREPBWT
H
T2_80).
HHNRM
MWT = Household
H
weight
w
after no
onresponse-aadjustment ffor the mail sscreener,
Approach
A
2 (rreplicate weigghts HHREP
PNRMWT1 – HHREPN
NRMWT80).
HHNRWT = Household
H
weight
w
after no
onresponse aadjustment fo
for the teleph
hone screenerr,
Approach
A
1, 2 (replicate weights
w
HHR
REPNRWT1 – HHREPN
NRWT80).
HHPSM
MWT = Final
F
poststraatified househ
hold weight ffor the mail sscreener, Approach 2
(rreplicate weigghts HHREP
PPSMWT1-H
HHREPPSM
MWT80).
HHPST
TWT = Final
F
poststraatified househ
hold weight ffor the telephhone screeneer, Approach
h 1,2
(rreplicate weigghts HHREP
PPSTWT1 – HHREPPST
TWT80).
WPERSBWT = Within-house
W
ehold person base weight,, Approach 11,2.
O
perso
on base weigh
ht = HHPST
TWT * WPER
RSBWT (rep
plicate weightts
PERSTBWT = Overall
PERSTREPB
P
BWT1-PERSTREPBWT880).
RSWT = Final
F
raked/trrimmed person weight, A
Approach 1,22 (replicate w
weights
FINPER
FINPERSWT
F
T1-FINPERSSWT80).
VICTIM
MWT = Victimization
V
n weight (repllicate weightss VICTIMW
WT1-VICTIM
MWT80).
INCIDENTWT= In
ncident weigght (replicate weights INC
CIDENTWT
T1-INCIDEN
NTWT80
NCVS-CS Pilot
P
Report
I-2
| File Type | application/pdf |
| File Title | Microsoft Word - NCVS-CS_Pilot_Report_Revised (031213).docx |
| Author | giambo_p |
| File Modified | 2014-03-13 |
| File Created | 2013-03-12 |