Attachment E

Attitudes Towards the Use of
Administrative Records
Jennifer Hunter Childs, Ryan
King, Casey Eggleston, Monica
Wroblewski
U.S. Census Bureau

Acknowledgements
•  Ryan King and Aleia Clark Fobia, U.S.
Census Bureau
•  Paul Scanlon, National Center for Health
Statistics
•  Michelle Smirnova, University of Missouri,
Kansas City

Research Questions
•  Under what circumstances do people support of the use
of administrative records for statistical purposes?
•  How does varying the source of an administrative record
effect a respondent’s attitude?
•  How does varying the social benefit of use of the record
effect the respondent’s attitude?
•  How does varying the monetary impact of the use effect
the respondent’s attitude?
•  What demographics and/or attitudes are related to a
respondent’s opinion of their use?

Data Collection
•  Gallup Nightly Survey
–  200 responses daily
–  Subsample of National RDD Sample
–  Landline and Cellphone

•  Initial Gallup contract was 25 questions collected
Feb. 2012
–  September 2013
•  19 question core, 6 rotating questions

•  Contract extension was 10 questions collected
October 2013 – March 2014
–  5 question core, 5 rotating questions
4

Administrative Records Modules
•  Module 1
–  Data collected for 25 weeks
•  23,511 total responses

•  Module 2
–  3 sets of questions
–  2 sets collected for 4 weeks
•  5,557 and 5,537 total responses

–  1 set collected for 6 weeks
•  8725 total responses

5

Module 1
•  In order to produce statistics, federal agencies can ask people for
information in a survey or get it from another source. If you knew
your name and other information would never be singled out and
would only be used for statistics, would you prefer that FSA’s:
•  Ask you for your employment information in a survey -- or ask a
state agency, like the employment or workforce office for it?
•  Ask you for information on your use of healthcare services in a
survey -- or ask your insurance company for it?
•  Ask you about the cost of products you buy in a survey -- or use
commercial records, like grocery store loyalty cards?
•  Ask you for your earnings history information from you in a survey
-- or ask the Social Security Administration for it?
•  Ask you for your income information from you in a survey -- or ask
the IRS for it?

6

Percent Reporting Preferring Data
From Administrative Record or Survey

Percent of Respondents Reporting that Fed
Gov shares a Common Database of Name,
Address and DOB

Module 2
•  Initial “Cold Ask”
•  Framed the question “why some people
like the idea of getting data from other
sources”:
–  Saving Government Money
–  Saving Respondent Time
–  Improving Accuracy
–  For Social Good

9

“Cold Ask”
•  The National Center for Health Statistics, or the NCHS,
collects information on people’s health and medical care
through a variety of surveys. The NCHS could obtain
some of your medical information from your health care
provider instead of asking you on a questionnaire. If you
knew that this information was only being obtained to
produce statistics, and that your personal information
would remain unavailable to the public, how would you
feel about the NCHS obtaining your medical information
from your health care provider? Would you be strongly
in favor of it, somewhat in favor of it, neither in favor nor
against it, somewhat against it, or strongly against it?
•  .

Frames
•  Now I will read you some benefits of the NCHS obtaining medical
information from health care providers. For each statement, please tell
me if these social benefits would make you strongly in favor of this….
•  SAVE MONEY - Getting information from healthcare providers would
help save some of the $100 million in government money that the NCHS
spends on surveys every year
•  SAVE TIME - Getting information from healthcare providers would reduce
the time and effort it takes people like you to participate in NCHS surveys
that require people to recall detailed medical histories, doctors’ visits, and
treatments
•  IMPROVE ACCURACY - Getting information from healthcare providers
would improve the accuracy of NCHS numbers, especially for people
who have difficulty accurately recalling their medical history.
•  SOCIAL GOOD - Getting information from healthcare providers would
help NCHS get a better idea of the health care needs of the nation which
would allow them to make better funding decisions for cancer research,
elderly medical care and hospitals

11

Experimental Design
Rota*on
 1

Rota*on
 2

Rota*on
 3

Reques*ng
 Agency

Census

NCHS

BLS

Record
 Source

SSA

Health
 Care
 Provider

Loyalty
 Card

Data

Name
 and
 Age

Medical
 Informa>on

Informa>on
 About

Purchases

12

Reported
 Preference
 for
 Use
 of
 Administra>ve
 Records
 in

"Cold
 Ask"
 Context

0.6
Census/Govt Record
NCHS/Health Record
BLS/3rd Party Record

0.49
0.5
0.43

0.41

0.4

0.4

0.39
0.35

0.3
0.19
0.2

0.19
0.16

0.1

0.0
Against

Neither

Favor

13

Relative Effectiveness of
Frames

Logistic Modeling Reported Views Towards
Use of Administrative Records
•  More Likely To Report Favor Use of Admin Records
–  All Rotations
•  Reporting trust in federal statistics
•  Reporting belief that agencies will keep data confidential

–  Census/BLS
•  Reporting that one can easily find out how federal statistics are
produced (Transparency)

–  NCHS
•  Reporting that policy makers need good statistics (Relevance)
•  Data users

–  BLS
•  Hispanic vs. White/Other (but not Hispanic vs. Black)

15

Logistic Modeling Reported Views Towards
Use of Administrative Records (2)
•  Less Likely To Report Favor Use of Admin
Records
–  All Rotations
•  Reporting little confidence in newspapers/
universities
•  Don’t Know/Refuse to report income

16

Reasons Given for Favoring Use of
Administrative Records
•  Census/Government Record
–  Save Money ~ 23%
–  Save Time ~ 19%
–  Improve Accuracy ~ 18%

•  NCHS/Health Record
–  Social Good ~ 29%
–  Improve Accuracy ~ 19%

•  BLS/Commercial Record
–  Social Good ~ 21%
–  Improve Accuracy ~ 21%

17

Reasons Given for Not Favoring
Use of Administrative Records
•  Census/Government Record
–  Privacy ~ 39%
–  Trust ~ 20%

•  NCHS/Health Record
–  Privacy ~ 56%
–  Trust ~ 20%

•  BLS/Commercial Record
–  Privacy ~ 53%
–  Trust ~ 20%

•  Few respondents (1-6%) mentioned accuracy as a
concern

18

Did the Shutdown Impact
Views towards Admin Records?
“Cold Ask”

Save Money Frame

100

100

80

80

60

60

40

40

20

20

0

Before
 Shutdown

Census

During
 Shutdown

NCHS

BLS

0

Before
 Shutdown

Census

During
 Shutdown

NCHS

BLS

19

Conclusions
•  Even with the observed increases in likeliness to favor
use of records when informed of cost savings and social
benefit, we still only observe nominal gains in favorability
and continue to see favorability hover around 50%.
•  Likeliness to support the use of administrative records
continues to correlate with overall trust in the
government.
•  Consistent with other independent research currently
going on at the Census Bureau, pleas that involve the
local social good seem to be more effective than
focusing on the national good.