Download:
pdf |
pdfAmerican Time Use Survey Web Data Collection
Final Report
July 2015
Prepared for:
Bureau of Labor Statistics
Washington, DC
Prepared by:
Westat
An Employee-Owned Research Corporation®
1600 Research Boulevard
Rockville, Maryland 20850-3129
(301) 251-1500
Table of Contents
Chapter
Page
General Introduction .........................................................................................
vii
Part I
Chapter
1
Introduction ........................................................................................................
1-1
2
The American Time Use Survey.......................................................................
2-1
3
Internet Coverage and Response to Web Surveys .........................................
3-1
4
Measurement of Time Use ................................................................................
4-1
5
Review of Surveys Measuring Time Use Over the Web...............................
5-1
6
Conclusion ...........................................................................................................
6-1
References ............................................................................................................
R-1
1
Design considerations for Web design ............................................................
4-3
2
Summary of time diary surveys reviewed ........................................................
5-2
Tables
Introduction
iii
Contents (continued)
Part II
Chapter
Page
1
Background..........................................................................................................
1-1
2
Allocation of Respondents to Mode ................................................................
2-1
3
Contact Procedures ............................................................................................
3-1
4
ATUS Diary Design for Web Collection ........................................................
4-1
5
Discussion ............................................................................................................
5-1
6
Scope of the Mock-up Task ..............................................................................
6-1
References ............................................................................................................
R-1
1
Contact procedures for mixed-mode design ..................................................
3-1
2
Internet use by household characteristics – American Community
Survey ...................................................................................................................
3-3
Distribution of ATUS tier 1 activity codes .....................................................
4-8
1
Example ATUS time diary ................................................................................
4-1
2
Summary of diary day.........................................................................................
4-3
3
Summary screens for work and income-generating activities ......................
4-4
4
Summary screen for secondary childcare ........................................................
4-5
Tables
3
Figures
Introduction
iv
Contents (continued)
Part III
Chapter
Page
Comparability Report .........................................................................................
1
Review of Proposed Mixed-Mode Design ........................................
1
Evaluation of Proposed Design ..........................................................
2
Additional Experimental Factors ........................................................
6
Adjustments for Mode Effects ...........................................................
7
Diary Design ..........................................................................................
7
References ............................................................................................................
R-1
1
Contact procedures for mixed-mode design ..................................................
1
2
Experimental design ...........................................................................................
4
Tables
Introduction
v
Part IV
Chapter
Page
Synthesis and Final Mock-Up ...........................................................................
1
Synthesis .................................................................................................
Mock-up .................................................................................................
1
2
References ............................................................................................................
R-1
1
Homepage for ATUS mock-up ........................................................................
3
2
Contextual overlay ..............................................................................................
4
3
Recording time spent on activity ......................................................................
5
4
Recording activities, who was with respondent, and location .....................
6
5
Checking for gaps in reported time use...........................................................
6
6
Identification of gaps in reported time use .....................................................
7
7
Example summary screen ..................................................................................
8
8
Example summary activity screen ....................................................................
8
Figures
Introduction
vi
Contents (continued)
Appendixes
Appendix
A
Paper Diary Examples .......................................................................................
A-1
B
Web Diary Examples .........................................................................................
B-1
C
Mobile Diary Examples .....................................................................................
C-1
D
Most Prevalent ATUS Activity Codes Using the Second Level of
ATUS Activity Codes .........................................................................................
D-1
E
Most Prevalent ATUS Activity Codes Using the Third Level of
ATUS Activity Codes .........................................................................................
E-1
Revised ATUS Activity Classification .............................................................
F-1
F
Introduction
vii
This page is intentionally blank
Introduction
General Introduction
This report presents research conducted by Westat about the feasibility of moving the American
Time Use Survey (ATUS) from telephone data collection to a mixed-mode design involving both
web and telephone modes. A move to a mixed-mode design has the potential to help ATUS
improve response and be prepared for the survey climate of the future.
This web design task focuses on the diary and follow-up summary questions (work, volunteer,
secondary childcare, and eldercare) and consists of three main components. First, the literature
review in Part I reviews the literature on web and mixed-mode data collection to help inform any
potential issues of introducing web data collection in ATUS. The literature review discusses findings
from the literature on coverage, response and measurement of web data collection instruments. We
also review several time use surveys that have considered or are using web data collection.
Next, the web design task in Part II provides recommendations on the design of web data collection
for the ATUS. The web design task considers allocation of respondents to mode in a mixed-mode
design, contact procedures, and question design considerations for a web instrument.
Part III includes a discussion of comparability issues between web and telephone data collection.
This section of the report covers methods to evaluate the proposed design including errors of
nonobservation (e.g. coverage and nonresponse error) and errors of observation (e.g. measurement
error). We also discuss additional experimental factors to consider, adjustments for mode effects,
and recommendations for evaluating elements of the web diary design.
Finally, Part IV synthesizes the material in the report and briefly describes a preliminary mock-up of
the recommended diary design.
Introduction
ix
Introduction
This page is intentionally blank
Introduction
x
Part I
Literature Review
Introduction
1
The goal of this literature review is to provide background to the Bureau of Labor Statistics (BLS)
about the use of web instruments to collect time use information. This will help to inform BLS
about whether the American Time Use Survey (ATUS) should move from CATI to a mixed-mode
design including both self-administered web and CATI. Other important goals are to improve
response and to be prepared for the survey climate of the future. Any movement from CATI to a
different mode could also introduce comparability issues with the ATUS historic data set, which at
present comprises 11 years of data. Hence this review will also investigate what the existing literature
on mode differences suggests on this topic. Importantly, the results of this literature review will help
inform the web design task that is to follow.
The literature review includes research from the more general mode literature involving web data
collection and where possible we include examples of web surveys used to collect time use diaries.
We begin with a brief review of the ATUS design and processes that are relevant for moving to
mixed-mode data collection. Next, we discuss response and coverage issues for web surveys. Third,
we discuss measurement issues for time diary surveys and the implications for mixed-mode data
collection. Then we detail some specific time use diary studies found in the literature search. Finally,
we end with some general conclusions from the literature review about the web design task.
Literature Review on ATUS Web Collection
1-1
This page is intentionally blank
The American Time Use Survey
2
ATUS is a unique source of information about the amount of time that Americans spend on a
diverse range of activities, including nonmarket activities like housework. The sample is drawn from
households that have completed their final round of the Current Population Survey (CPS). Hence,
the ATUS is a rich frame design that includes much information about the respondent that could be
used to contact the respondent using different modes of data collection. A designated person (DP)
from the CPS panel, age 15 and older, is selected and assigned a specific day of the week to report
for. Ideally, any mode that is used to collect time use information should be able to encourage
respondents to complete the diary on the designated day. The respondent reports for a 24-hour time
diary about “yesterday.” The interview is conducted using CATI. There are 17 major activities in the
ATUS lexicon. A diary requires at least five activities to be included in the published figures. The
median number of activities per diary is about 20-21. Data are published annually.
The ATUS interview consists of several parts: (1) information about household composition, (2) the
respondent’s labor force and school enrollment status, (3) the 24-hour time diary about “yesterday,”
and (4) summary questions about specific topics of interest. “Yesterday” is defined as 4 a.m. the
previous day to 4 a.m. on the day of the interview. All activities undertaken during the time period
are reported on. The smallest reporting period is 1 minute. Respondents can report the duration of
the activity or the start and stop times. The ATUS interview lasts approximately 15 minutes; there
are sometimes optional modular questions that usually run approximately 5 minutes. Modules are
funded by other agencies and relate to time use.
The ATUS interviewers have a number of follow-up probes that they use to gather more
information on specific topics and responses in the diary. One type of follow-up probe is
automated; with this probe, a pop-up screen in the CATI instrument displays the probe question
text. Any activity that lasts more than 2 hours, with the exception of sleeping, working, or attending
high school, requires the interviewer to verify the activity by asking the question: Did you stop
(activity) to do anything else at that time? Work activities that last more than 4 hours generate an
instrument pop-up requiring the interviewer to ask the respondent if he/she took any breaks of 15
minutes or longer. If a respondent reports sleeping more than 10 hours, the interviewer must verify
the duration of the sleep activity. Another automated probe is for respondents who do not report
any eating or drinking activity in the diary. Interviewers must verify that the respondent did not eat
Literature Review on ATUS Web Collection
2-1
The American Time Use Survey
2
or drink on the diary day. There are a number of other probes that do not automatically pop up but
which interviewers must still ask about certain activities to collect the level of detail required for
activity coding. The interviewer is instructed to record the main activity only. If a respondent reports
simultaneous activities, the interviewer asks questions to see if the respondent can separate the
activities, if possible; if not, the interviewer asks the respondent to identify which activity was the
main activity.
A number of summary questions are asked after the time diary to collect information on specific
topics of interest. This includes secondary child care, or care for children under age 13 that is done
while doing an activity other than primary child care; activities done for pay; volunteer activities; and
eldercare. Eldercare is defined as care provided in response to a “condition related to aging.”
The inclusion of these different types of questions in the ATUS raises important design
considerations for any mode change. The comparability of the new mode with CATI data collection
is important since the ATUS has been using this single mode for more than 10 years. For example,
the follow-up probes help to ensure that appropriately detailed information is collected in the ATUS
and that data quality is sufficient. It is important to understand if any new mode of data collection is
able to replicate this capability to reduce mode differences in the resulting data.
Literature Review on ATUS Web Collection
2-2
Internet Coverage and Response
to Web Surveys
3
These next two sections explore error properties of web surveys to provide an understanding of the
implications for total survey error when moving to web data collection. We focus this review on the
most relevant error sources for the ATUS: coverage, nonresponse, and measurement. We exclude a
discussion of sampling error since the ATUS has an existing sampling frame consisting of outgoing
participants of the CPS. We begin with a discussion of coverage and response.
Internet Coverage
One important consideration in migrating to a mixed-mode design involving web surveys is the
extent to which the U.S. population is covered by Internet access. Surveys disagree on the exact
percentage of households with Internet access; however, the number of households that have access
to the Internet has clearly been growing over the past several years. Tourangeau, Conrad, and
Couper (2013) summarized estimates of coverage from several different surveys. They show that,
according to the CPS, between 1997 and 2009, Internet coverage among U.S. households has grown
from 18 percent to 69 percent. A Pew survey in 2011 suggested that coverage may be as high as 78
percent. As Tourangeau, Conrad, and Couper indicate, obtaining an exact percentage of households
with access is difficult since individuals may be able to obtain access in multiple locations including
home, work, school, or other public places such as the library. In addition, the advent of mobile
devices such as smartphones and tablets is also probably improving coverage.
In addition to the overall level of coverage, it is also important to understand whether this coverage
differs across demographic groups. Tourangeau, Conrad, and Couper (2013) present evidence from
the Health Information National Trends Survey (HINTS) conducted in 2005 and 2007 that indeed
there are differences. The most glaring differences occur with age, education, and race. As one might
expect, Internet coverage declines with age. HINTS 2007 shows that 80 percent between the ages of
18 and 34 are covered, whereas only about 45 percent of those 65-74 are covered. In addition,
approximately 80 percent of those with some college are covered, but only about 56 percent of
those with a high school diploma are covered. Finally, 75 percent of non-Hispanic whites are
covered, but only about 56 percent of non-Hispanic blacks, and 49 percent of Hispanics are
covered. It remains to be seen how quickly these differences will disappear as coverage of the entire
Literature Review on ATUS Web Collection
3-1
Internet Coverage and Response to Web Surveys
3
population increases. These differences do suggest that different demographic groups are likely to
respond quite differently to the introduction of a web option in a mixed-mode design for the ATUS.
Overall Response to Web Surveys
In order to understand how people respond to web surveys we first consider response to standalone web surveys and how the level of response compares to other modes. Meta-analytic studies
that have compared response rates from web surveys to other survey modes have found that web
surveys achieve lower response rates than the other modes. Two studies have found that on average
the response rate for web surveys is approximately 11 percent lower than other modes of data
collection (Lozar Manfreda et al., 2008; Shih and Fan, 2008). Most of these studies have compared
web surveys to mail surveys. Relatively little is known about how response rates for web surveys
compared specifically with telephone surveys. A meta-analysis by Lozar Manfreda (2008) included
only five telephone studies, but found that response rates for web surveys are about 12 percentage
points lower than their telephone counterparts. A similar difference was found between mail and
web.
There is also evidence that break-offs are greater for web surveys than for CATI surveys. CATI
surveys have the advantage that the interviewer can call the respondent back and complete the
interview, thus having an interruption as opposed to a termination (McGonagle 2013). Peytchev
(2009) identifies the following characteristics of the questionnaire that lead to higher break-offs in
web surveys: (1) questions that are longer and more complex, that is, more burdensome for the
respondent, have higher break-off rates. (2) Open-ended questions lead to higher break-off rates. (3)
The presence of a progress indicator that makes the survey appear long can also contribute to a
higher break-off rate.
Factors Affecting Participation
The available evidence suggests that many of the same strategies for obtaining participation in
traditional survey modes such as mail, phone, or in-person may also apply to web surveys. A few
studies suggest that prenotification is effective for web surveys (Crawford et al., 2004; Harmon,
Westin, and Levin, 2005). Mailed advance letters have been found to be more effective than advance
e-mails. In terms of the actual invitation to complete the survey, there are some reasons to believe
that e-mail invitations might be better. One would expect that the ability to include a link in the eLiterature Review on ATUS Web Collection
3-2
Internet Coverage and Response to Web Surveys
3
mail to go directly to the survey would provide an advantage; however, this ease of access has to be
weighed against the likelihood that the sample person will actually open the e-mail. Tourangeau,
Conrad, and Couper (2013) discuss how the prevalence of spam and the volume of e-mail create an
environment where the typical survey request may not even be noticed. In addition, even slight
errors in e-mail addresses will cause the invitation to be returned, which places a high value on
having correct information on the sampling frame. In contrast, there is more tolerance for error with
postal mail addresses. That being said, it is likely that the preference for e-mail or postal invitations
may vary by population. Kaplowitz et al. (2012) found that a postcard invitation was better for
faculty and staff, but not students from a university. There are some design choices that may
improve response in e-mail invitations. For example, personalization of the e-mail tends to increase
response, especially when the sender of the e-mail has high status (Heerwegh et al., 2005; Joinson
and Reips, 2007; Joinson, Woodley, and Reips, 2007).
The number of contacts with the respondent also influences response. Increases in the number of
contact attempts may be particularly intriguing by web since each additional contact attempt incurs
no monetary cost. Meta-analyses have shown that there are larger differences between modes when
more contact attempts are made (Lozar Manfreda et al, 2008; Shih and Fan, 2008). That is, although
each additional contact attempt on the web may recruit more respondents, there seems to be a larger
diminishing effect relative to other modes. Tourangeau, Conrad, and Couper (2013) suggest that
continued reminders by e-mail may harden respondents to the request. Furthermore, these
reminders may also be less likely to be read than they are in other modes.
Incentives are another common method for attempting to improve response. There are two
approaches to incentives in the broader survey methodological literature that have the most success.
Prepaid incentives achieve higher response than promised incentives, and offering cash achieves
higher response than all other forms of incentives (Singer, 2002). In spite of these findings, the web
survey literature on incentives is replete with examples of lottery incentives conditional on
completion of the survey (Goritz, 2006). This also appears to be true in time diary surveys involving
web data collection (Bonke and Fallesen, 2010; Wu et al., 2012; see also Table 2 this document).
Bonke and Fallesen (2010) present model-based evidence that a combination of lottery incentives
improved response in a mixed-mode survey involving a time diary. Goritz (2006) presents evidence
that lotteries are helpful for improving response in commercial panels, but there is scant research
showing that these types of incentives are effective in other settings. The notion of a lottery
incentive with capped costs is appealing for containing the cost of a survey. In addition, it is often
difficult to deliver prepaid cash incentives in the context of a web survey. There are examples of
efforts to mimic this transaction with electronic means such as PayPal that have been unsuccessful
Literature Review on ATUS Web Collection
3-3
Internet Coverage and Response to Web Surveys
3
(e.g. Bosnjak and Tuten, 2002). It is important to note that the ATUS may not have as many
limitations as most web surveys do because of its rich sampling frame. For example, prepaid
incentives could be mailed to households in an advance letter to the sample person.
Mixed-Mode Considerations
Due to the lower coverage and response to web surveys compared to other modes, it is
understandable that using the web alone may not be the best approach. Therefore, it may be better
to use the web in combination with other modes of data collection. Mixed-mode data collection
aims to improve overall response and to shift as many respondents as possible to the web, because it
is a cheaper mode and each additional respondent does not add much cost to data collection (unless
a cash incentive is involved).
There are different strategies that can be undertaken to increase the number of respondents who
complete the survey via web. One approach is a concurrent mixed-mode design where the
respondent is simultaneously offered the option of completing the interview using one of two or
more modes. There is a gathering consensus in the literature that offering multiple modes
simultaneously may actually hinder response. The exact reason for this lower response is unknown
at this time. One explanation is that offering a choice adds complexity and burden to the survey
request by requiring respondents to weigh the pros and cons of each mode, with the result that
some respondents opt not to participate at all (Medway and Fulton, 2012). Offering a choice also
provides the opportunity for respondents to procrastinate about the decision long enough that they
lose the materials needed to participate (Tourangeau, Conrad, and Couper, 2013). Another
possibility is the incompatibility of two modes. For example, respondents may be contacted in one
mode, but prefer to complete the survey in another mode; this increases the time and effort needed
to participate and they decide not to respond (Tourangeau, Conrad, and Couper, 2013). Finally, it is
possible that some respondents may actually start completing the survey in a mode such as web and
then encounter implementation failure or technical problems. As a consequence the respondent may
have lost the materials required for the other mode (e.g., mail) or may decide to drop the effort to
participate all together as consequence of the problems encountered (Tourangeau, Conrad, and
Couper, 2013; Medway and Fulton, 2012).
Another approach to mixed-mode design is the sequential mixed-mode design. In a sequential
mixed-mode design all respondents are given the opportunity to complete the survey in a single
mode initially, typically an inexpensive mode, but are followed up in another mode. Bonke and
Literature Review on ATUS Web Collection
3-4
Internet Coverage and Response to Web Surveys
3
Fallesen (2010) present an example of sequential mixed-mode design in the context of time diary
research. They sent sample persons a letter offering the choice of a 10-15 minute telephone
interview or a web survey with an access code. The respondents were called and asked to complete
the survey over the telephone if the web questionnaire had not been received within two weeks.
They achieved a similar level of response to a previous paper time use survey of the Danish
population in 2001. The design led to about one fifth of the respondents completing the time diary
over the web.
The sequencing of modes to improve response and reduce costs is an active area of research that is
currently unsettled. A study by Millar and Dillman (2011) presents evidence that sequential strategies
that push the web mode more strongly in earlier contacts can achieve similar response rates and lead
to a higher proportion of respondents completing the survey over the web compared to designs that
stress other modes (e.g., mail) early and web later. Others have failed to replicate these findings. In
addition, Cantor et al. (2010) conducted a mode experiment in a 2-phase survey. An ABS sample
was sent a paper screener to identify Veterans and their families. The Veterans and their families
identified in phase 1 were given a choice of responding to the phase 2 questionnaire via mail or web.
The authors found that the response rate for the mail-only design was higher than the sequential
mixed-mode design. Hence, there is still much research that needs to be conducted to determine the
optimal sequencing of modes.
Literature Review on ATUS Web Collection
3-5
This page is intentionally blank
Measurement of Time Use
4
Introducing another mode of data collection can introduce differences in data that may be due to
differences in the measurement properties of the available modes. There are two sources for
measurement differences between modes when measuring time use. First, the survey methodological
literature has identified various characteristics of modes that influence measurement error. This
literature suggests that characteristics of the mode such as the presence of the interviewer,
computerization, and the channels of communication that are utilized may influence the
measurement properties of a survey mode. Second, there are survey design decisions that are made
in the context of time use diaries that may have implications for mode differences if the decision
leads to a preference for one or the other modes. These decisions include whether the activities are
recorded verbatim or in precoded categories, whether the time period is fixed or variable, whether
secondary activities are recorded, whether contextual questions are asked, differences in recall
periods, whether probes or follow-up questions are needed, and the number of diary days that are
measured (Gershuny, 2011).
General Mode Differences between Web and Telephone
There are some important differences between web and telephone surveys that may lead to
differences in the data that are collected. One key difference is the presence of the interviewer. An
interviewer can have both negative and positive influences. On the positive side, the interviewer can
be helpful for recruiting respondents to participate. The interviewer also can be trained to identify
problems and to probe appropriately. This is particularly true in the ATUS where flexible
interviewing is used to aid recall in the time diary (BLS, 2013). The introduction of web raises the
question of whether recall can be improved in other ways with web data collection. For example, it
may be possible that web data collection allows the respondent to spend more time on the
instrument, which aids recall. The influence of web on recall also depends on the timing of diary
completion. It is possible that the web diaries could be completed sooner than is possible for a
telephone interview such that recall is improved. On the negative side an interviewer creates more
social presence, which leads to less reporting of socially undesirable behavior and more reporting of
socially desirable behavior (Tourangeau and Yan, 2007). The introduction of a web data collection
Literature Review on ATUS Web Collection
4-1
Measurement of Time Use
4
has the potential to improve reporting of socially desirable or undesirable behaviors relative to
CATI, which may lead to differences in data quality between the two modes.
Computerization is another difference that might lead to differences between modes, because the
approach to computerization affects the complexity of the survey instrument. The current ATUS
instrument includes some complicated features such as controlling the overall flow of the instrument
and probes that are used to ensure data quality (e.g., when the respondent makes a potential error
such as when an activity takes longer than two hours or forgets to enter a travel activity). Existing
time use diaries completed over the web can control question flow and have shown that probes can
be programmed to check for data quality (Houle, 2014; Wu et al., 2012).
Web introduces a visual channel of communication to the ATUS, whereas the current instrument is
completely auditory. This may be the source of the biggest difference between the existing CATI
instrument and any potential web instrument. The visual display of information in web data
collection creates many possibilities and potential challenges. Since the respondent is now interacting
directly with the instrument, the time diary will have to be designed in a way that the respondent can
complete the interview with little or no assistance. Tourangeau, Couper, and Conrad (2013) review a
vast literature on visual design of web surveys that would be impossible to review thoroughly in this
literature review. The design of a web instrument will have to consider the specific layout of
information on the screen and its comparability with the ATUS CATI data. Much of the visual
layout will be driven by the specific decisions about measuring time use so we will mention visual
design implications during the following discussion. The most effective way to evaluate the effect of
the visual layout is through rigorous usability testing.
It is also noteworthy that another feature of the web that could be beneficial to ATUS is
interactivity. Tourangeau, Conrad, and Couper (2013) discuss a number of interactive features of the
web and guidelines for using these features. Some of these features can introduce bias and require a
great deal of caution. For example, images can have unintended impact on the answers (Couper,
Tourangeau, and Kenyon, 2004). There are at least two features that may be of assistance to ATUS.
First, the instrument can be made more interactive by including such features as showing
respondents a summary of their time use activity as it is reported or at the end of the interview so
that respondents can review what they have reported. The Modular Online Time Use Survey
(MOTUS) demo allows for this capability. Another feature is the introduction of video to describe
to respondents how they can complete the diary and also provide them with background on the
purpose and uses of the ATUS. Some of the web and mobile applications that we reviewed utilized
this type of interactivity (Minnen et al., 2014; Sonck and Fernee, 2013).
Literature Review on ATUS Web Collection
4-2
Measurement of Time Use
4
Aspects of Time Use Measurement that May Affect Mode Differences
Next, we briefly review a few key issues that may influence potential differences with the existing
CATI instrument. We base this discussion on a number of sources that discuss some of the critical
design decisions related to time use measurement (e.g. Gershuny, 2011; United Nations, 2013). The
topics we will discuss are shown in Table 1.
Table 1.
Design considerations for web design
Design choice
Examples
Current ATUS design
Recording of activities
Verbatim/Precoded
Verbatim (coded by interviewer)
Time period
Fixed/Variable
Variable
Number of activities
Primary/Secondary
Primary (except childcare)
Additional questions
Who were you with? What was your
location?
Who was in the room with you?
Who accompanied you? Who was
with you? (The WHO question
varies depending on the activity
reported.) Where were you while
you were [ACTIVITY]?
Probing
Missing travel time, missing sleep time,
missing eating and drinking
Multiple probes included
Reporting period
Yesterday/Tomorrow
Yesterday
Diary days
1 – 7 days
1 designated day
One design issue that will affect both the design of a web instrument and may be a cause of mode
differences is how the activities are recorded in the diary. The time use literature refers to “full
diaries” where the respondent’s activities are recorded more or less verbatim and “light diaries”
where respondents would record their activity using a precoded list. Full diaries also often collect
more information, but the key difference seems to be the entry of the actual activity. A light diary
approach may reduce burden, but may not measure time use as accurately as a full diary. The current
ATUS is a full diary that begins with the interviewer asking, “Yesterday at 4:00 AM, what were you
doing?/what did you do next?” so that respondents can verbalize in their own words what they were
doing. The interviewer enters the activity information. The ATUS CATI instrument includes
precodes for a few straightforward activities; on the other hand, most activities are entered verbatim
and coded in a separate coding step. The ATUS lexicon includes 17 major activity classifications and
more than 400 detailed activity classifications.
There are examples of paper self-administered diaries taking both “full” and “light” approaches. For
example, New Zealand uses a full approach by having respondents enter the activity in their own
Literature Review on ATUS Web Collection
4-3
Measurement of Time Use
4
words on each line of the diary. Japan and Finland have tried both full and light approaches (see
Appendix A for examples). The Finnish experience is that light diaries produce less time in activities,
such as housework (Paakkonen and Vaisanen, 2012). The web diary instruments that we reviewed all
used some kind of precoded category scheme that varied in complexity. In addition, there are
different ways to display and have the respondent select the codes. Drop-down boxes and radio
buttons were used to select activities depending on the complexity of the list. Some diaries such as
the MOTUS demo version allows for drop-down selection or keyword search. The Dutch Time Use
Survey (DTUS) smartphone pilot allowed for open-ended entry if the respondent’s activity did not
fit one of the existing categories (Sonck and Fernee, 2013).
Another important design decision is whether the time period in which information is entered is
fixed or variable. For example, activities can be entered in 10 or 15 minute intervals or the
respondent can simply be asked how long an activity lasted or report the beginning and end time of
the activity. The current ATUS asks “How long did you spend [ACTIVITY]?” so it is a variable time
period. The New Zealand paper time diary presents fixed time periods of five minutes (see
Appendix A) and the DTUS smartphone pilot presents fixed ten-minute intervals. Intuitively,
variable time periods probably represent the most accurate measure of reported time use since
respondents can say exactly how long they participated in an activity. There are examples of web
diaries that utilize variable time periods like the ATUS including the MOTUS demo and Canadian
web pilot diary (see Appendix B).
Diaries can also differ in the extent to which they allow respondents to enter simultaneous activities.
For example, some diaries allow the respondent to only enter primary activities, whereas others
allow respondents to enter primary and secondary activities. For the most part, the current ATUS
only allows the respondent to enter primary activities. The only exception to this is that the ATUS
asks respondents to go back over the day and report about childcare and eldercare activities.
Time diaries often ask additional questions related to the context of time use. For example, the
current ATUS asks questions such as, “Who was in the room with you?” and “Where were you
while you were [ACTIVITY]?” These types of additional questions can be incorporated into a web
instrument easily if the questions are common for each reported activity. Even if the additional
questions are conditional on the type of activity that is reported, the web instrument should be able
to incorporate flows and skips that ask these contextual questions, but it may require more screens
to be programmed per reported activity (Houle, 2014). The Canadian Time Use Survey web pilot
has maintained a multiple screen approach to incorporate these flows.
Literature Review on ATUS Web Collection
4-4
Measurement of Time Use
4
Another issue to consider is whether the instrument needs to include any probes or follow-up
questions to ensure data quality. For example, the current ATUS asks probes about any activity that
lasts more than 2 hours, with the exception of sleeping, working, or attending high school. It
requires the interviewer to verify the activity by asking the question: Did you stop (activity) to do
anything else at that time? Work activities that last more than 4 hours generate an instrument pop-up
requiring the interviewer to ask the respondent if he/she took any breaks of 15 minutes or longer. If
a respondent reports sleeping more than 10 hours, the interviewer must verify the duration of the
sleep activity. Another automated probe is for respondents who do not report any eating or drinking
activity in the diary so that interviewers must verify that the respondent did not eat or drink on the
diary day. ATUS interviewers are also trained to capture sufficient detail for a number of other
activities; these probes are not automated. It is unclear to what extent existing web time use diaries
utilize these types of probes; however, it seems likely that web instruments would be able to perform
data quality checks like those in the current CATI instruments.
Another potential difference that could be introduced between the existing ATUS and a web diary is
the reporting period. The literature describes the use of “yesterday” and “tomorrow” diaries.
Yesterday diaries have respondents construct how they spent their time during the previous 24
hours. Tomorrow diaries involve the respondent recording his or her activities during a pre-specified
day (United Nations, 2013). The current Eurostat guidelines advocate tomorrow diaries (Eurostat,
2009). The current ATUS asks about the previous 24 hours starting at 4 AM the previous day to 4
AM the day on which the respondent is called. It may make more sense to approach the respondent
in advance about filling out the diary on a pre-specified day. This would mean the diary would be
open for the respondent to report on that day. Respondents could enter the activity at any time
during the day shortening the actual recall period. A decision would have to be made about how
long the diary should be left open for reporting. This deviation in reporting periods could be a
source of mode differences; however, the available research suggests that yesterday and tomorrow
diaries produce similar estimates (Gershuny 1995; Robinson and Godbey, 1999).
The ATUS currently allows for respondents to answer about their time use on a specific diary day in
order to adequately represent time across days of the week. This can also be easily replicated on the
web, but it is worth mentioning that web diary collection has the additional benefit of being able to
easily and without monetary cost increase the number of diary days. All respondents, or just a
subsample of respondents could be asked to complete the additional diary days. On the other hand,
increasing the number of diary days would increase response burden for the respondent and could
have deleterious effects on response rates.
Literature Review on ATUS Web Collection
4-5
Measurement of Time Use
4
Mobile Devices
The introduction of mobile devices has complicated the collection of data over the web somewhat.
There may be relatively minimal impact for some devices such as tablets; however, the limited screen
space available can make it difficult to perform some survey tasks (Peytchev and Hill, 2010). For
example, grids may be difficult or impossible to implement on a small screen. Some have developed
apps that can be used to measure time use on smartphone devices (e.g. Sonck and Fernee, 2013).
One needs to consider that some respondents will attempt to complete the diary over their phone as
these devices become more popular. Although these devices present certain challenges, they also
present opportunities for time use measurement. For example, respondents carry their smartphones
with them most of the time. Hence this makes it possible to record time use in more or less real
time, and also makes it possible to send notifications to respondents to implement experience
sampling techniques (Sonck and Fernee, 2013).
Data Quality
Several of the studies we reviewed provided information about the quality of the data collected over
the web versus more traditional modes such as paper or telephone. Most of these studies present
evidence that time use surveys conducted online over the web or through the use of mobile apps
can produce data of similar quality to existing modes of data collection; however, the literature
would benefit from more experimental comparisons to draw stronger conclusions about the
comparability of data quality across modes.
Data quality in time use studies is often measured by metrics such as number of activities reported,
the sequence of the activities reported, the number of simultaneous activities reported, or the
completeness of the data. Bonke and Fallesen (2010) report a higher number of activities and
sequences reported using the web-based instrument compared to telephone and conclude that the
web-based instrument produces better data quality. These findings occurred in the context of a
mixed-mode design rather than an experimental comparison; however, the differences still occurred
even after controlling for socio-demographic characteristics such as gender, age, education, and
labor market status. A pilot of 51 families in the UK Millennium Study found that there were
21 weekday episodes reported for a web diary versus 19 for paper and 16 for a mobile app
(Chatzitheochari, 2014). No significance tests were reported for the study so it is not possible to tell
whether these results are statistically significant. This same study found that the number of full
productive diaries was similar for diaries completed online over the web, through an app, or on
Literature Review on ATUS Web Collection
4-6
Measurement of Time Use
4
paper. Sonck and Fernee (2013) report that the DTUS smartphone pilot found that the number of
activities reported using the smartphone app was similar to the number of activities reported on
previous surveys of the Dutch population. On average, respondents reported 27 episodes in the
pilot test of the app in 2011, which was more than was reported in the 2005 DTUS (24) but fewer
than was reported in the 2006 DTUS (34). Both the 2005 and 2006 surveys were collected using
paper-administered diaries.
The Study of Use of Products and Exposure Related Behavior (SUPERB) suggests that information
of similar quality can be collected by web and telephone diaries. This study involved several contacts
over an 18 month period. Everyone first completed one weekday diary and one weekend day diary
over the telephone. The respondents completed one weekday diary and one weekend day diary per
quarter online for the remainder of the study period. Less than one percent of both the web and
phone diaries were incomplete (Wu et al., 2012). Some differences were found between the
telephone and web diaries. For example, the first year telephone diaries reported a significantly
higher number of location-activity records compared to the web diaries. Other differences were
corrected by design changes. First, there were more instances of implausible sleep hours (e.g. <3
hours or > 16 hours) entered in the web diaries compared to the telephone diaries. Some of these
were caused by data entry errors by the respondent, but the initial diary also did not allow the user to
edit reported information. In addition to allowing the respondent to edit information, adding a
warning message decreased the differences in reported sleep between the telephone and web diaries.
Second, the web diaries included less time spent in transit and more missing transit records. A
change to the web diary that added a warning if the respondent failed to enter transit between stops
resulted in fewer missing transit records so that there were no longer significant differences between
the telephone and web diaries.
Literature Review on ATUS Web Collection
4-7
This page is intentionally blank
Review of Surveys Measuring Time
Use Over the Web
5
We reviewed several studies that have used self-administered web-based diary collection methods. There
are a number of countries that are considering web-based applications for time-use surveys (United
Nations Economic Commission for Europe, 2010). The available literature provides information on a
number of pilot studies or development efforts that are underway to incorporate web-based data
collection. These efforts are at various stages in the development process. We found many of these
articles by using search engines such as Google or Google scholar. We also reviewed the references of
articles that we found. Our goal for this literature review was to collect as much information as possible
about the entire survey process including contact procedures, use of incentives, and instrument design,
but many of the reviewed surveys do not have complete documentation. In addition, the published
literature is currently sparsely populated with examples of web time use diaries. Hence, much of the
research we reviewed is in conference presentations or other documents that were obtained either online
through Internet searches or through communication with the study directors. Table 2 reviews the
information that we learned about each survey.
Overall Designs of the Reviewed Studies
There are a number of different types of studies that we reviewed using different study designs
depending on the research goal. Researchers from Belgium using the Modular Online Time Use Survey
(MOTUS) tool have conducted a time use survey over the web alongside a paper and pencil diary study
in order to benchmark the data collected over the web with a traditional data collection method (Joeri et
al., 2013). Researchers from the Netherlands undertook a feasibility study to understand the use of
smartphones to collect time use data (Sonck and Fernee, 2013; Fernee and Sonck, 2013; Fernee and
Scherpenzeel, 2013). Other studies utilized a mixed-mode approach that combined web data collection
with another more traditional mode such as telephone or paper by either offering respondents an option
to switch to the web or transferring everyone to the web (Bonke and Fallesen, 2010; Wu et al., 2012;
Chatzitheochari, 2014). The type of online survey may also vary within these mixed-mode designs. While
most of the studies used traditional web instruments or mobile apps, Japan conducted their Survey on
Time Use and Leisure Activities in 2011 and included an option to complete an electronic questionnaire
instead of a paper questionnaire. The electronic questionnaire was delivered as a downloadable
spreadsheet.
Literature Review on ATUS Web Collection
5-1
This page is intentionally blank.
Literature Review on ATUS Web Collection
Table 2.
Summary of time diary surveys reviewed
Survey
Modular
Online Time
Use Survey
(MOTUS)-Belgium
Dutch Time
Use Survey
(DTUS)
Canadian
Time Use
Survey
Modes
Web
Study type
Pilot
App
Pilot
Letter with
instructions
Web
Pilot
NA
UK Millenium
Cohort Study
Web, App,
Paper
Pilot
NA, Time use
paper for
memory aid
when Internet
access not
available.
Contact
procedures
Letter with
link to
website and
instructions
Incentives
Lottery
Nonresponse
follow-up
E-mail, SMS
Response
NA
Input
Precoded
list,
keyword
search
Context
Location,
Others
present,
transport
mode
Others
present
Probes
NA
Sample
40,000 adults
(ages 18-75) from
the Belgium
population register
NA
Probes
for
travel
activity
150 adults from a
market research
panel
For actual
production fielding,
one person age
15+ will be
selected per
household from a
telephone frame.
The telephone
frame is generated
from the Canadian
Census and other
government
administrative
records.
Teens age 14 from
Wave 6 of the
Millenium Cohort
Study. 50 teen in
the pilot test and
100 teens in the
dress rehearsal.
Notifications
through app
NA
Precoded
list, open
NA
NA
Precoded
list
Location,
Others
present,
NA
Text
reminders
70%
participation
from existing
panel
Precoded
list
Location,
Others
present,
satisfaction
5-2
Panel
reward
points
NA
Level of detail
of activities
225 precoded
activities
Other
information
Special
modules can
be added
41 precoded
activity
categories
72 activity
codes that roll
up to 14
activity
categories
GPS,
experience
sampling
Use of
technology,
Mood
44 activity
codes that
rolled up to 12
broad activity
groupings
NA
Literature Review on ATUS Web Collection
5-3
Table 2.
Summary of time diary surveys reviewed (continued)
Survey
Danish Time
Use and
Consumption
Survey
Modes
Web,
Telephone,
mixed-mode
Study type
Pilot
Study of Use
of Products
and Exposure
Related
Behavior
(SUPERB)-California
Web,
Telephone
One time
study
Web survey
accessed
through study
website with
log-in
Lottery
Japanese
Survey on
Time Use and
Leisure
Activities
Paper,
electronic
questionnaire
Production
Part of
sample
offered option
of
downloadable
questionnaire
NA
Contact
procedures
Letter with
choice of
telephone or
online
Incentives
Lottery
Nonresponse
follow-up
Telephone
reminders
with no online
complete
Response
48%; 20% of
completes
online
Input
Precoded
list
Context
NA
Probes
NA
NA
16% for
parents of
young
children, 49%
older adults
Precoded
list
Location
Probes
for
travel
and
sleep
activity
NA
300/10,700
completes
via electronic
questionnaire
NA
NA
NA
Sample
Adults ages 18-74
randomly selected
from the
administrative
records held at
Statistics
Denmark. The final
sample size was
6,000 completed
interviews.
186 households in
CA with a child age
8 or younger.
Sample randomly
drawn from birth
certificates of
children born
between 20002005. 64 adults
ages 55 or older
randomly selected
from tax assessors'
housing units
records.
Multi-stage
nationally
representative
sample of
households in
Japan. Household
members of age 10
and over completed
the survey.
Level of detail
of activities
"Pre-coded
system" with
key word
search.
Other
information
NA
No discussion
of codes or
open field entry
NA
Part of sample
completed a
pre-coded form
and the other
part of the
sample
completed a
free entry for.
NA
Review of Surveys Measuring Time Use Over the Web
5
Contact Procedures
A variety of information is provided to respondents in the process of contacting households for
participation in a Web survey. It was typical for studies to send an introductory letter to the
respondents. The content of the letter varied across the projects that we reviewed. In addition to
stressing the importance of the survey and mentioning incentives, some included other motivational
techniques. For example, the pilot for the MOTUS in Belgium informed the selected household that
they would receive an overview of their time use and a comparison of their own time use with the
remainder of the population (Joeri et al., 2013). In addition, the MOTUS provided links to the
website where the respondent could find an introductory video on the web page that stressed the
societal importance of the survey. Some studies talked specifically about the log in procedures for
completing the web diary. For example, Wu et al. (2012) describe the log in procedures for the
SUPERB study, which required participants to access the web survey through a study website with a
unique ID and password. The Danish Time Use and Consumption Survey utilized a mixed-mode
design that included an introductory letter offering the sampled household the choice of completing
a telephone interview lasting 10-15 minutes or completion of a questionnaire on the web with an
access code (Bonke and Falleson, 2010).
Instructional materials were also commonly sent to sample households to help respondents
understand how to use the web tool to record their time use. The amount and nature of the material
varied across studies. The Dutch Time Use Survey (DTUS) tested the use of a smartphone app
(Fernee and Sonck, 2013; Fernee and Scherpenzeel, 2013; Sonck and Fernee, 2013). The participants
were recruited from an online panel and were sent several materials in addition to an introductory
letter. For example, the participants were sent a paper manual describing how the app worked. The
participants could also access a YouTube video that demonstrated the use of the app. This study
also provided smartphones to participants who did not own their own smartphone so there was also
a YouTube video that explained how to use the smartphone. Helpdesk support was also available
for respondents who encountered difficulty with the app. The UK Millennium study gave 51
respondents recruited from wave 6 of a longitudinal study the choice of completing a diary for two
diary days either over the web or through a mobile app (Chatzitheochari, 2014). A paper diary was
offered to those who did not have Internet access. The respondents were provided mode-specific
leaflets with instructions.
Studies also included detailed information and materials to support the completion of the time use
diary. For example, the self-administered task of entering time use categories requires the
respondent to be familiar with the coding scheme. Hence, the DTUS provided a paper list of time
Literature Review on ATUS Web Collection
5-4
Review of Surveys Measuring Time Use Over the Web
5
use activities (Sonck and Fernee, 2013). The UK Millennium Cohort Study collected data through an
app and also from web users. Both groups were provided a paper time use notebook to keep a
record of their activities in case they could not access the diary on the allocated day
(Chatzitheochari, 2014).
Incentives
Incentives were common for participating in the time use surveys that were conducted over the web.
The type of incentives used differed from the typical cash incentives. Lotteries were often used to
encourage sample households to participate in the survey. The MOTUS pilot in Belgium used a
lottery consisting of one hundred prizes ranging from 10-500 Euros raffled each month among
those who completed the surveys (Joeri et al, 2013). SUPERB offered an unspecified incentive and
entered respondents in a raffle to encourage retention in the study which required completing
several time use diaries over an 18 month period (Wu et al., 2012). The DTUS smartphone pilot
recruited respondents from a panel and used loyalty points for the panel as an incentive (Sonck and
Fernee, 2013). There was no experiment to determine whether these incentives were effective
or not.
The Danish time use and consumption survey used incentives in the context of a mixed-mode
design. The survey required the sample adult and other household members to complete a
demographic questionnaire, time use survey, and a consumption booklet about the expenditures by
members of the household. They also utilized lottery incentives. For the first six months of the
survey, respondents that were interviewed participated in a lottery with three prizes (one prize of
5000 Danish krone, two prizes of 1000 Danish krone) if they filled out the questionnaire, diary, and
booklet. This lottery was announced in a letter that was sent out in the consumption booklet, which
was mailed after completion of the questionnaire and diary. After six months the lottery prizes were
increased (one prize each of 10,000, 5,000, and 2000 Danish krone). In addition, the prizes were
doubled if the respondents completed the instruments on the web. Finally, cinema tickets to all
family members were offered if they filled out the remaining non-completed instruments. Bonke and
Falleson (2010) present model-based evidence that the use of these incentives increased the
response rate by 10 percent.
Literature Review on ATUS Web Collection
5-5
Review of Surveys Measuring Time Use Over the Web
5
Nonresponse Follow-up
An advantage of electronic data collection is the possibility of sending reminders to participants who
have not completed the survey. The MOTUS pilot asked respondents to provide their e-mail
address and mobile telephone numbers when they logged on to the website for the first time (Joeri
et al, 2013). This gave the study an opportunity to send customized e-mail messages or SMS
messages to respondents who started their diary but did not finish. The diaries were completed
continuously throughout the day with the peak reporting times being 8 am and 10 pm. Sonck and
Fernee (2013) report that for the DTUS “If no activities were recorded in the smartphone diary for
two hours, a notification was sent to respondents to remind them about completing the diary up to
the most recent activity” (p. 26). The UK Millennium Cohort Study sent text reminders about
completing and submitting the diary (Chatzitheochari, 2014).
Reminders using more traditional modes of data collection may also be used with web data
collection. The Danish time use and consumption survey used telephone follow-ups at the
beginning of the survey period in its mixed-mode design (Bonke and Fallesen, 2010). The sample of
adults was drawn from a register maintained by Statistics Denmark. Selected adults were sent a letter
offering them the choice to complete the survey either by telephone or web. Respondents who had
not completed the survey after one week using the web instrument were given a telephone reminder
to complete the survey and were also offered the option of completing the interview on the
telephone. There were up to nine reminder calls per selected person.
Response
We found response rates reported for two of the surveys that were reviewed. The UK Millennium
Study reported an overall participation rate of 70 percent. The participation rate was 75 percent for
the first diary day and 65 percent for the second diary day (Chatzitheochari, 2014). As shown in
Table 2, the sample for this study consisted of participants from an existing panel. The mixed-mode
(telephone and web) Danish time use and consumption survey reported a 48 percent response rate
for the time diary portion of the survey, which was similar to the 49 percent response rate achieved
by the Dutch time use survey using paper only (Bonke and Fallesen, 2010). Approximately one in
five respondents completed the time use diary over the web. The Japanese found an even lower
proportion of respondents choosing a downloadable electronic questionnaire over a traditional
paper diary. Only 300 out of 10,700 respondents chose the electronic questionnaire. Wu et al. (2012)
found that 16 percent of parents and children completed all questionnaires across their entire 18
Literature Review on ATUS Web Collection
5-6
Review of Surveys Measuring Time Use Over the Web
5
month study, whereas 49 percent of older adults completed all questionnaires. This difference most
likely reflects the increased burden for parents of young children since they had to fill out multiple
surveys rather than any bias in preference for the mode.
Input Type
Accurately recording time use can be a challenge in self-administered diaries. In contrast to the
current ATUS where an interviewer who is trained and familiar with the time use coding scheme,
respondents using self-administered modes may have difficulty understanding what exactly to record
in the diary. However, many self-administered instruments use a similar style coding scheme that the
respondents can use to categorize their time use activity. The images of the diaries shown in the
appendixes provide detailed examples of how these coding schemes work in practices. We briefly
describe some of these approaches in this literature review.
In general, most of the surveys we reviewed used some kind of hierarchical list for the respondent to
code their time use into. Many of these designs are based on some variant of a “light diary” where a
reduced set of time use categories captures the most relevant time use activities; however, the exact
codes vary from study to study (Pääkkönen and Väisänen, 2012). For example, in the web diary that
was developed for the Canadian Time Use Survey, respondents reported what they were doing using
72 “Light diary” codes that roll up to 14 high level activities (Houle, 2014). Respondents were then
asked how long the activity lasted. Ten-minute time intervals were used. The Canadian instrument
also collected up to two pre-defined simultaneous activities per main activity. The organization of
the categories can have an important impact on what actually gets recorded by the respondent. For
example, evaluation of the Canadian Time Use Survey revealed that including travel as a category
under all major headings helped to reduce the number of times travel was omitted. The number of
pages used to collect this information varied over the development of the diary. The initial sevenpage diary design utilized mostly radio buttons to select categories and allowed for flows, dynamic
text, and edits, but this design was viewed as burdensome. The instrument was reduced to a twopage design that utilized drop-down boxes and also allowed for flows and edits. Developers found a
one-page design to be insufficient since it did not allow for flows and edits.
Both the web instrument and the app in the UK Millennium Cohort Study included a hierarchical
list of 44 activities under 12 top-level groupings. The app is designed for a question by question
format beginning with what were you doing, where were you, who were you with, and how much
did you like [activity]. After each activity is entered the respondent can see everything that has been
Literature Review on ATUS Web Collection
5-7
Review of Surveys Measuring Time Use Over the Web
5
entered up to that point. In contrast, the web instrument is presented as a time grid
(Chatzitheochari, 2014).
The other surveys reviewed also included some type of precoded hierarchical list. Even within this
type of presentation, there may be different options for enabling respondents to code time use
activities. The MOTUS allows the respondent to either point-and-click on the list to select the
activity or search for keywords that might guide the respondent to the correct activity (Joeri et al.,
2013). For example, a respondent could select “personal care” and then “sleep” or the respondent
could just type in “sleep” using the key word search and find the appropriate category that way. The
mobile app developed for the DTUS included a precoded hierarchical list, but also allowed users to
type in their activity in an open-ended field (Sonck and Fernee, 2013).
Contextual Questions
All of the surveys we reviewed collected answers to contextual questions in addition to time use.
The MOTUS prototype included questions about the location of the respondent, who the
respondent was with and the transportation mode when transportation was selected (Joeri et al,
2013). The UK Millennium cohort study collected how much the respondent liked an activity in
addition to where the respondent was located and who they were with (Chatzitheochari, 2014). The
SUPERB study asked whether the activity occurred indoors, outdoors, or in a motorized vehicle, in
addition to where the respondent was located (Wu et al., 2012). The DTUS collected whether the
respondent was alone or with someone that he or she knew (Sonck and Fernee, 2013). There were
plans for the Canadian Time Use Survey to measure whether or not the respondent used a tablet,
smartphone, computer or laptop, how pleasant an activity was, where the respondent was located,
and who the respondent was with in their Fall 2014 pilot survey (Houle, 2014).
Probes, Editing, and Follow-up Questions
Web data collection shares the advantages of computerization with CATI data collection. This has
two implications for the ATUS. First, it means that the two modes can incorporate skip patterns or
appropriate follow-ups conditional on the type of time use activity that the respondent reports. This
was a design element that motivated the Canadians to consider multiple-page versions of their web
instrument that would allow for flows, dynamic text, and edits (Houle, 2014). Second, the web
instrument can mimic some of the probing and editing behavior of the CATI instrument. Our
Literature Review on ATUS Web Collection
5-8
Review of Surveys Measuring Time Use Over the Web
5
review of time use studies using web data collection suggests that this is an important aspect to
consider in terms of data quality. The SUPERB study found that including follow-ups about travel
and sleep activity helped to eliminate differences in data quality between telephone and web data
collection (Wu et al., 2012). The Canadian Time Use Survey also planned to probe for missing travel
activity in their pilot study when different locations are entered without any travel in between
(Houle, 2014).
Other Information Collected
The use of mobile devices to collect time use data may have some particular advantages to collecting
supplemental data for time use purposes. The DTUS smartphone pilot used the GPS feature of the
smartphone to track location information of respondents who gave their consent to collect such
information (Sonck and Fernee, 2012; Fernee and Scherpenzeel, 2013). The app also allowed for the
possibility of experience sampling throughout the day so that researchers could understand how
respondents were feeling at certain times of the day and then correlate this information with time
use activity. For example, the app would send a notification to the respondent at particular times of
the day and ask the respondent to report his or her mood.
Summary
There are a few lessons that can be learned from this review of time diary surveys using web data
collection methods. First, web data collection has been explored using both single- and mixed-mode
designs. Second, it is common for introductory material to include detailed instructions and support
concerning how to complete the web survey. Third, there is very little guidance about the level of
response to web time diary surveys from the existing literature; however, the mixed-mode design
utilized by Bonke and Fallesen (2010) demonstrates that the response rate can be similar to previous
time use studies using a single mode. Bonke and Fallesen also found that only about one in five
respondents chose to complete their diary over the web. Fourth, lottery incentives were the most
common type of incentive used for web time diary surveys. Bonke and Fallesen (2010) provide some
evidence that these incentives work, but there is currently a dearth of evidence from the existing
literature to support this finding. Also, regarding response, it is noteworthy that nonresponse followups and reminders are relatively easy with web data collection once the requisite contact information
(e.g., email addresses) has been collected. Finally, there are many aspects of web data collection that
need to be evaluated to improve comparability between web data collection and telephone
Literature Review on ATUS Web Collection
5-9
Review of Surveys Measuring Time Use Over the Web
5
collection. For example, simpler coding categories will probably be necessary for self-administration.
Contextual questions, follow-ups, or probes can be asked in much the same way as on the
telephone, but the use of these condition questions will need to be considered specifically in the
design since it may influence the number of screens that may be needed for flows and edits.
Lessons Learned From Other Related Research
Future Mobility Survey. The Future Mobility Survey (FMS) is a travel survey that used a GPS app
loaded on to the respondent’s smartphone to track the respondent’s travel. The GPS data were
uploaded to a central server and the server mapped and analyzed the GPS data. The respondent was
then prompted in an on-line recall task to provide detail about the processed GPS data. The FMS
found that the battery life of a smartphone was a significant challenge since the GPS app demanded
much of the smartphone’s battery capability. The FMS also found that the web interface for the
recall task needs to be as simple and intuitive as possible since there is no interviewer to assist the
respondent with the recall task.
Web Mode Test for the Consumer Expenditure Diary Survey. At present, the CE production
diary is paper-based. In 2013, BLS conducted a 3-month feasibility test of a web CE diary. In the
main, the current version of the CE web Diary functioned well and met the objectives originally
defined by BLS. On the other hand, the current version of the web Diary is not ready for
production. Findings from the test of the CE web Diary include the following findings:
Response rates were lower for the web Diary than for the corresponding rates in the
production paper-based diary.
There were also overall fewer expenditures (median of 13 fewer expenditures) reported
for the web Diary than for the production paper-based diary.
On the other hand, the web Diary respondents reported higher median expenditure
amounts for certain sections and exhibited lower item nonresponse than the CE
production paper-based diary.
The web diary test demonstrated the importance of flexible and easy usernames and
passwords, since the usernames and passwords proved problematic for a number of
respondents.
Web diary respondents should be allowed to continue reporting a day or so after the
final day of data collection.
Web diary paradata should be examined on a daily basis to identify respondents who
have not logged on and to note the number of days since last logon.
Literature Review on ATUS Web Collection
5-10
Review of Surveys Measuring Time Use Over the Web
5
Life in a Day App. The National Cancer Institute (NCI) along with Westat developed the Life in a
Day mobile phone diary-based app for capturing time spent in daily activities. With a goal of
improving the quality of self-reported physical activity behavior, the app was designed to allow
subjects to report their daily activities either in real time or retrospectively.
NCI followed best practice principles in developing the app and did not provide a tutorial for
respondents to learn how the app worked. In usability testing, the app proved not to be sufficiently
simple or intuitive to learn. Respondents had trouble leaning some of the basic features of the app
and others never discovered some of the basic features. This again points to the extreme importance
of making the app simple and self-evident.
The usability testing also showed that the registration process for the app was highly burdensome
and could deter some respondents from participating in the survey. Further, the app allows
respondents to customize activity buttons. But the customization process proved so burdensome
that some respondents chose to use the pre-designated buttons, thus providing less detailed
information. The timer feature was not easy to use and some respondents found that the times
recorded for their activities were inaccurate.
Literature Review on ATUS Web Collection
5-11
Conclusion
6
This literature review was undertaken to inform the design of a web instrument that could be used
in a mixed-mode design in the ATUS. We briefly reviewed the current features of the ATUS
relevant for the design. We also reviewed the literature illustrating potential sources for error related
to coverage, nonresponse, and measurement error for a web instrument in a mixed-mode design for
the ATUS. Finally, we reviewed several surveys that are using web instruments to measure time use.
The literature suggests that including web in a mixed-mode design is feasible. Other countries have
implemented a similar approach (e.g., Bonke and Fallesen, 2012); however, there are a number of
issues that will need to be addressed with the introduction of a new mode in the ATUS. Nearly all of
the studies we reviewed are pilot studies that have not been implemented in a production
environment. Therefore time use studies with extensive experience with web collection are still rare.
Many of the issues that need to be addressed concern how closely the web can measure time use in
way that is consistent with the current design using telephone data collection. The current literature
offers some guidance about how self-administered data collection may relate to the current
telephone administered design, but there are many areas that need further research. For example, the
data will need to be evaluated to understand if socially desirable or undesirable behaviors are
reported at different rates between the web and telephone modes. Likewise, there will need to be
research about whether a light diary approach with precoded categories can capture data similar to
the current ATUS activity data detail. In addition, the current mode research is still coalescing
around the best strategy for pushing respondents to complete the survey on the web. Hence, ATUS
will want to carefully consider different options for combining the two modes together.
Literature Review on ATUS Web Collection
6-1
This page is intentionally blank
References
Bonke, J., & Fallesen, P. (2010). The impact of incentives and interview methods on response
quantity and quality in diary-and booklet-based surveys. In Survey Research Methods, 4 (2),
91-101.
Bosnjak, M., & Tuten, T.L. (2002). Prepaid and promised incentives in Web surveys – an
experiment. Social Science Computer Review, 21, 208-217.
Bureau of Labor Statistics. (2013). American Time Use Survey User’s Guide. Washington D.C.:
Bureau of Labor Statistics.
Cantor, D., Brick, P.D., Han, D., & Aponte, M. (2010). Incorporating a web option in a two-phase
mail survey. Paper presented at the 60th Annual Conference of the American Association for
Public Opinion Research., Miami Beach, FL, May.
Chatzitheochari, S. (2014). Measuring teenage time use in the UK Millennium Cohort Study: A
mixed-mode approach. Paper presented at the IATUR Conference., Turku, Finland, July.
Couper, M. P., Tourangeau, R., & Kenyon, K. (2004). Picture this! Exploring visual effects in web
surveys. Public Opinion Quarterly, 68, 255-266.
Crawford, S.D., McCabe, S.E., Saltz, B., Boyd, C.J., Freisthler, B., & Paschall, M.J. (2004). Gaining
respondent cooperation in college web-based alcohol surveys: Findings from experiments at
two universities. Paper presented at the 59th Annual Conference of the American Association
for Public Opinion Research., Phoenix, AZ, May.
Eurostat (2009). Harmonised European Time Use Surveys: 2008 guidelines. Luxembourg: Office
for Official Publications of the European Communities.
Fernee, H. & Scherpenzeel, A.C. (2013). The smartphone in survey research: experiments for time
use data. The Survey Statistician, 19-25.
Fernee, H. & Sonck, N. (2013). Is everyone able to Use a smartphone in survey research? Tests
with a time-use app with experienced and inexperienced users. Survey Practice 6(4). Retrieved from
http://www.surveypractice.org/index.php/SurveyPractice/article/view/245/html.
Gershuny, J. (1995). “Time budget research in Europe”. Statistics in Transition 2 (4), 529–551.
Gershuny, J. (2011). Time-Use Surveys and the Measurement of National Well-Being. Centre for
Time-use Research. Department of Sociology, University of Oxford.
Goritz, A.S. (2006). Incentives in web studies: Methodological issues and a review. International
Journal of Science, 1, 58-70.
Literature Review on ATUS Web Collection
R-1
References
Harmon, M.A., Westin, E.C., & Levin, K.Y. (2005). Does type of prenotification affect web survey
response rates? Paper presented at the 60th Annual Conference of the American Association
for Public Opinion Research., Miami Beach, FL, May.
Heerwegh, D. (2005). Effects of personal salutations in e-mail invitations to participate in a web
survey. Public Opinion Quarterly, 69, 588-598.
Houle, P. (2014). Innovations in measuring time use in Canada: Developing a self-administered
electronic diary. Paper presented at the IATUR Conference., Turku, Finland, July.
Joeri, M., Ignace, G., van Tienoven, T. P., & Djiwo, W. (2013). MOTUS: Modular online time-use
survey. NTTS 2013 - New Techniques and Technologies for Statistics. Retrieved from
http://www.cros-portal.eu/sites/default/files/NTTS2013fullPaper_184.pdf.
Joinson, A.N., & Reips, U.D. (2007). Personalized salutation, power of sender and response rates to
We-based surveys. Computers in Human Behavior, 23, 1372-1383.
Joinson, A.N., Woodley, A., & Reips. U.D. (2007). Personalization, authentication, and selfdisclosure in self-administered Internet surveys. Computers in Human Behavior, 23, 275-285.
Kaplowitz, M.D., Lupi, F., Couper, M.P., & Thorp, L. (2012). The effect of invitation design on web
survey response rates. Social Science Computer Review, 30, 339-349.
Lozar Manfreda, K., Bosnjak, M., Berzelak, J., Haas, I., & Vehovar V. (2008). Web surveys versus
other survey modes: A meta-analysis comparing response rates. International Journal of Market
Research, 50, 79-104.
McGonagle, K. (2013). Survey breakoffs in a computer-assisted telephone interview. Survey
Research Methods, 7(2), 79-90.
Medway, R. L., & Fulton, J. (2012). When more gets you less: A meta-analysis of the effect of
concurrent web options on mail survey response rates. Public Opinion Quarterly, 76(4),
733-746. Retrieved from http://poq.oxfordjournals.org/content/76/4/733.full.pdf+html.
Millar, M. M., and Dillman, D. A. (2011). Improving response to Web and mixed-mode surveys
Internet using address-based sampling and mail contact procedures. Public Opinion Quarterly, 75,
249-269.
Minnen, J., Glorieux, I, van Tienoven, T., Daniels, S., Weenas, D., Deyaert, J., Van den Bogaert, S.,
Rymenants, S. (2014). Modular Online Time Use Survey (MOTUS) – Translating an existing
method in the 21st century. International Journal of Time Use Research, 2014, 73-93.
Pääkkönen, H. and P. Väisänen (2012). Comparison of results from light diaries and full-scale diaries
in the Finnish time use survey. Statistics Finland (unpublished).
Peytchev, A. (2009). Survey breakoff. Public Opinion Quarterly, 73, 74-97.
Robinson, J.P., G. Godbey (1999). Time for Life. The Surprising Ways Americans Use Their Time.
Second edition. University Park: The Pennsylvania State University Press.
Literature Review on ATUS Web Collection
R-2
References
Shih, T.H., & Fan, X. (2008). Comparing response rates from web and mail surveys: A metaanalysis. Field Methods, 20, 249-271.
Singer, E. (2002). The use of incentives to reduce nonresponse in household surveys. In R.M.
Groves, D. Dillman, J.L. Eltinge, & R.J.A. Little (Eds.), Survey Nonresponse (pp. 163-177).
New York: John Wiley.
Sonck, N. & Fernee, H. (2013). Using smartphones in survey research: a multifunctional tool
Implementation of a time use app; a feasibility study. The Netherlands Institute for Social
Research/SCP, The Hague.
Tourangeau, R., Conrad, F.G., & Couper M.P. (2013). The Science of Web Surveys. New York: Oxford.
Tourangeau, R. & Yan, T. (2007). Sensitive questions in surveys. Psychological Bulletin, 133,
859-883.
United Nations Economic Commission for Europe Conference of European Statisticians (2013).
Guidelines for harmonizing time use surveys (Meeting Minutes). Third Meeting of the
2012/2013 Bureau. Luxembourg, 5-6 February 2013.
Wu, X., Bennett, D.H., Lee, K., Cassady, D.L., Ritz, B., & Hertz-Picciotto, I. (2012). Feasibility of
using web surveys to collect time-activity data. Journal of Exposure Science and
Environmental Epidemiology, 2012, 116-125.
Literature Review on ATUS Web Collection
R-3
This page is intentionally blank
Part II
Design Task
Background
1
The American Time Use Survey (ATUS) is a unique source of information about the amount of
time that Americans spend on a diverse range of activities including nonmarket activities like
housework. The sample is drawn from households that have completed the final round of the
Current population Survey (CPS). A designated person (DP) from the CPS panel, age 15 and older,
is selected and assigned a specific day of the week to report for. The respondent reports for a
24-hour time diary about “yesterday.” The interview is conducted using computer-assisted telephone
interviewing (CATI).
The ATUS interview includes questions about (1) the composition of the household, (2) the
respondent’s labor force and school enrollment status (3) the 24-hour time diary about “yesterday”
and (4) summary questions about specific topics of interest.
The Bureau of Labor Statistics is interested in pursuing a mixed-mode design that includes the
collection of data via a web instrument. This document outlines a design proposal for implementing
a mixed-mode approach in ATUS consisting of both web and CATI data collection. Three design
elements are covered in this proposal:
Allocation of respondents to mode: A sequential mixed-mode design is proposed to
maximize the use of the web mode.
Contact procedures for a mixed-mode design: Recruiting designated persons to
participate in the ATUS will require some modifications to the contact procedures that
are currently employed on ATUS. Either advanced letters or e-mails, if available, may be
used to contact the designated person. The designated person can be switched to
telephone data collection if efforts to recruit via the web are unsuccessful.
ATUS Diary design for web collection: The ATUS instrument includes a number of
features that may need to be adapted or modified for web data collection including the
activity classification, level of detail that is collected, follow-up questions, clean-up
probes, and the collection of secondary activities such as child care or eldercare.
ATUS Web Design Task
1-1
This page is intentionally blank
Allocation of Respondents to Mode
2
The current ATUS design conducts all interviews by CATI. Designated persons from households
that provided a telephone number to the Census Bureau and completed the eighth interview of the
CPS are called to complete the interview over the telephone. Each respondent is assigned a
designated diary day and is called on the day after their designated diary day. Designated persons
from households that did not provide a telephone number are asked to call the telephone center on
a designated day to complete the ATUS interview. The designated person has up to eight weeks to
complete the interview on their designated day of the week. Throughout this proposal we assume
that ATUS will continue using the designated day approach.
One of the main goals of any mixed-mode design for the ATUS using both web and CATI modes
should be to maximize the number of participants who complete the diary online. Maximizing
response via the web would help to minimize the cost of the mixed-mode design since the marginal
cost for adding a CATI respondent is much higher than the marginal cost of adding an additional
web respondent.
Researchers have tried both concurrent and sequential mixed mode designs. In the concurrent
mixed-mode design, the designated person would be simultaneously offered the option of
completing the interview using either web or CATI. There is a gathering consensus in the literature
that offering a choice of modes may actually hinder response. Although several hypotheses have
been offered about why this is the case, the exact reason for this lower response is unknown at this
time (Medway and Fulton, 2012). For example, it is possible that offering two modes simultaneously
adds complexity and burden to the decision to participate.
Another approach is a sequential mixed-mode design where all respondents are given the
opportunity to complete the survey in a single mode initially, typically an inexpensive mode, but are
followed up in another mode. A study by Millar and Dillman (2011) presents evidence that
sequential strategies that push the web mode more strongly in earlier contacts can achieve similar
response rates and lead to a higher proportion of respondents completing the survey over the web
compared to designs that stress other modes (e.g., mail) early and web later. Bonke and Fallesen
(2010) present an example of a mixed-mode design in the context of time diary research that is not
strictly sequential, but encourages the use of the web mode first. Sample households in their study
ATUS Web Design Task
2-1
Allocation of Respondents to Mode
2
were asked to first complete a background question. Next, they were asked to complete two forms
for daily time use – one for a weekday and one for a weekend day. Finally, sample households were
asked to complete a booklet about their previous month’s spending. Sample households were sent a
letter offering the choice of a 10-15 minute telephone interview or a web survey with an access code.
The sample households were given one week to complete the interview over the web. Households
were then called and asked to complete the survey over the telephone if the web questionnaire had
not been received within one week. They achieved a similar level of response to a previous time use
survey of the Danish population in 2001 conducted on paper by mail. The design led to about one
fifth of the respondents completing the time diary over the web. The Bonke and Falleson study
design also included lottery incentives that were higher when the interviews were completed over the
Web rather than by telephone. They estimated the average treatment effect of the incentives in this
increased the response rate by ten percent.
We recommend that the ATUS follow a sequential mixed-mode design that pushes the designated
person to complete the interview via the web first and then switches to CATI data collection after a
pre-specified number of attempts has resulted in a non-interview. Ideally, information from the CPS
frame could be used to help allocate designated persons between CATI and web modes. For
example, ideally the CPS frame would include e-mail addresses and information about web access.
However, this type of information is not asked about in the current CPS interview. The CPS
interview does include information that is predictive of web access such as age, household income,
and education. This information could be used to tailor contact strategies for those who do not
respond to the initial request to complete the diary on the web. We will now present procedures for
contacting respondents using a mixed-mode design.
ATUS Web Design Task
2-2
Contact Procedures
3
Under the current ATUS design, the ATUS designated person is sent an advance mailing to notify
them that they have been selected for the ATUS sample. The mailing contains a letter printed in
both English and Spanish. The purpose of the letter is to explain the nature of the survey, notify the
designated person of day of their interview, and provide ATUS telephone, e-mail and Internet
contact information. The letter that is sent to households without telephone numbers is also
accompanied by a $40 inactive ATM debit card. The mailing also includes a brochure that provides
answers to some frequently asked questions about the survey.
Table 1 shows a proposed pattern of contacts for a mixed mode design. This design varies
depending on the week that the case is in the sample, whether or not a telephone number is available
for the household, and the likelihood of the designated person being a web user. The diary day in
Table 1 refers to the actual day that the respondent is to complete the diary about.
Table 1.
Wee
k
1
Contact procedures for mixed-mode design
Day
Telephone HH
High Web Use
Low Web Use
Non-telephone HH
7 days before
diary day
1 day after diary
day (if not
complete)
1 day after diary
day (if not
complete)
Advance mailing
Advance mailing
Advance mailing
Replacement letter
Replacement letter
Replacement letter
Replacement letter
Replacement letter
3
1 day after diary
day (if not
complete)
4-8
1 day after diary
day (If not
complete)
Switch to phone for
remaining weeks;
calls begin on the
day after the diary
day on week four
Interviewer calls
for interview on
day after diary day
Switch to phone for
remaining weeks;
calls begin on the
day after the diary
day on week three
Interviewer calls for
interview on day
after diary day
2
ATUS Web Design Task
3-1
Interviewer calls for
interview on day
after diary day
Switch to phone for
remaining weeks;
Send letter with prepaid calling card and
call-in number
Respondent calls in
for interview on day
after diary day
Contact Procedures
3
Many of the existing ATUS materials would still be helpful for the mixed-mode design in Table 1
with some modifications. First, the advance mailing should contain a letter and a brochure similar to
what is provided in the current ATUS advance mailing. The letter would have to be altered in order
to provide the designated person with a URL, and a user name and password or similar login
procedure to use once they have reached the ATUS diary website. The exact protocols for delivering
the login information provided may vary depending on the security requirements for BLS surveys.
We also recommend including a $2 prepaid incentive in the advanced mailing as a token of
appreciation for completing the survey. The purpose for the incentive is to encourage the designated
person to complete the interview via the web, which should lead to cost savings versus completing
the interview by telephone. It may also be helpful to provide materials that will help the respondent
complete the diary. For example, a copy of the activity categories that the respondent will be using
when filling out the web diary could be provided.
As shown in Table 1, a replacement letter would be sent to designated persons who did not
complete the interview during the first week. This letter would contain much of the same
information that is included in the current ATUS refusal conversion letter. It would also include the
log-in information and URL that the respondent uses to complete the web diary.
After week one, the design in Table 1 differs depending on whether the household provided a
telephone number during the CPS and the designated person’s likelihood of using the web. First, the
non-telephone households would have up to three weeks to complete the diary via the web. They
would receive the advance mailing plus two replacement mailings if needed. The designated person
would then be sent a letter with a $40 pre-paid debit card asking them to call-in to the telephone
center if they have not completed the interview by week 3. In other words, contact procedures for
the non-telephone households become very similar to the current ATUS contact procedures for this
group after week three. The initial web alternative gives non-telephone households another viable
option to complete the survey to increase response for this group.
The contact protocol for households with a telephone number varies depending on what is known
about the designated person’s likelihood of using the web from the CPS. Table 2 shows a summary
of Internet use for households in 2013 by age of the householder, educational attainment of the
householder, and household income. The table shows substantial drops in Internet use for
households where the householder is 65 or older, the householder has a high school degree or less,
and where the household income is less than $25,000.
ATUS Web Design Task
3-2
Contact Procedures
Table 2.
3
Internet use by household characteristics – American Community Survey
Household characteristics
Total households
N
Percent with Internet subscription
116,291
74.4
Age of householder
15-34
35-44
45-64
65 years or older
22,331
20,745
46,015
27,201
77.7
82.5
78.7
58.3
Educational attainment of householder
Less than high school graduate
High school graduate (or equivalent)
Some college or associates degree
Bachelor’s degree or higher
12,855
28,277
34,218
36,349
43.8
62.9
79.2
90.1
Household income
Less than $25,000
$25,000-$49,999
$50,000-$99,999
$100,000-$149,999
$150,000 or more
27,605
27,805
34,644
14,750
11,487
48.4
69.0
84.9
92.7
94.9
Source. Computer and Internet Use in the United States: 2013 – American Community Survey Reports.
Based on the evidence in Table 2, we suggest making fewer contacts to households where the
designated person is age 65 years or older, has less than a high school degree, or lives in a household
with less than $25,000 in income. We call these designated persons our “low web use” group. These
designated persons with a telephone number are switched to the telephone mode one week earlier
than others with a telephone number. Although the data in Table 2 refer to the householder, we
recommend using the actual age and education information for the designated person when making
the assignment to these groups.
The proposed design in Table 1 assumes that the designated person’s e-mail information is not
available from the CPS as is currently the scenario. Collecting the designated person’s e-mail address
at the time of the CPS would make invitations and follow-ups easier and less costly. Hence, we
would recommend collecting e-mail addresses in the CPS if possible. Even with e-mail addresses
available, we would still recommend an advance mailing that provides the respondent with the initial
request and information about the survey. Initial contacts by e-mail have the potential to be deleted
by e-mail filters or missed by the designated person. However, e-mail follow-up reminders and
follow-ups have two major advantages. First, there is less lag time between when the information is
sent and when the respondent receives the information. Second, direct links or log in information
for the web diary can be provided in the e-mail making it easier to open the diary. In addition, e-mail
follow-ups would be cheaper relative to any telephone follow-ups.
ATUS Web Design Task
3-3
This page is intentionally blank
ATUS Diary Design for Web Collection
4
Survey Questions and Overall Diary Design
Much of the general structure of the time diary from the current ATUS interview can be maintained
in the move to a web diary. Figure 1 illustrates a potential layout for collecting time diary
information. There are several features of the design that should enhance collection of time diary
information and minimize differences with the current ATUS design. First, the existing survey
questions for the time diary can be used with minimal changes. Figure 1 illustrates that the first
question a respondent would see on the screen once they entered the time diary is “Please select or
enter a description of your main activity on [DAY], [DATE] at [TIME].” This first question focuses
the respondent on their designated diary day. It is possible to collect paradata that indicates when
the respondent entered the diary information. It may also be important to confirm the respondent
answered about their designated diary day with a debriefing question at the end of the survey. The
time should be updated automatically as the respondent enters activities throughout the day.
Figure 1.
Example ATUS time diary
ATUS Web Design Task
4-1
ATUS Diary Design for Web Collection
4
The respondent could click a link to see a list of the full ATUS activity classification to help them
determine how to best classify an activity. In addition to using the dropdown list, we recommend
that the respondents have the opportunity to search for activities that they cannot find in the
dropdown lists. If the search is not successful, the activity can be recorded verbatim. These activities
could be coded using similar procedures for coding activities in the CATI diary. The object of the
search capability is to limit the amount of open ended comments that are necessary to code.
The web instrument would use progressive disclosure of the dropdown lists and follow-up
questions. For example, second tier (e.g., working at main job) categories would only appear after
the respondent had selected a first tier activity (e.g., work). Third tier categories could also be
utilized where more detail is needed. Likewise, the follow-up questions “How long did you spend on
this activity?”, “Who was with you during this activity?”, and “Where were you during this activity?”
would only appear after the activity was selected. This is useful since the follow-up questions are
dependent on selection of the activity. For example, the “who” and “where” follow-ups are not
asked when the respondent selects sleeping or some other personal care activities.
Respondents would have the choice of entering the time that they stopped the activity or how long
they spent on the activity, which is consistent with the current ATUS interview where the
interviewer can do either. The “Who” list would be populated from the household roster that was
asked earlier in the interview and also include important pre-coded persons who would not be part
of the roster and are part of the current ATUS “Who” list. The “Where” dropdown list could match
the existing list on the ATUS. A list of transportation modes would be displayed if the respondent
selects transportation as an activity.
One final important feature of the time diary is a table that summarizes the respondent’s time use as
it is entered throughout the day. This table provides a memory cue and context to the respondent as
they think about the next activity that they did during the diary day. This table would also provide a
logical place for the respondent to edit or add activities that they think about as they are filling out
the diary, but did not remember when they were first entering information in the diary.
Summary of Diary Day
At the end of the interview the respondent should be given the opportunity to review the activities
they reported for the entire day. Figure 2 gives an example of how this screen may appear. This
screen allows the respondent to see their day as a whole and is a logical place for the respondent to
ATUS Web Design Task
4-2
ATUS Diary Design for Web Collection
4
make any edits to their day. We could include as much information on this screen as necessary. For
example, we could include columns that display “Who” and “Where” information related to each
activity. The respondent may feel the need to edit after reviewing their day in its entirety. They may
also edit in response to pre-programmed quality checks for the respondent to edit the screen. For
example, the respondent might not have reported any eating and drinking during the day and the
instrument could probe the respondent about this scenario.
Figure 2.
Summary of diary day
Other Summary Questions
There are four sections of the ATUS interview that involve a summary of activity that may have
already been reported in the diary. These sections include work and income-generating activities,
secondary childcare, volunteering, and eldercare. We expect that by this time in the interview the
respondent has finalized their activity for the day so that it would not be necessary to return to the
diary to edit information. We now outline a strategy for collecting this information through a series
ATUS Web Design Task
4-3
ATUS Diary Design for Web Collection
4
of summary screens once the respondent has completed the diary. It may not be necessary to
include some summary screens (e.g. work and volunteer activity) since respondents are self-reporting
activity in the diary; however, for the purposes of the mock-up exercise it may be worthwhile to see
examples of how these screens might work in practice.
Work and income-generating activities. The current ATUS interview asks the following:
“We are interested in measuring the amount of time people spend working
both inside and outside their usual workplace. You said that you were
working from [start and stop time of worked at main job from time diary].
Were there any [other] activities that were done as part of your job or
business? Please do not include getting ready for work or commuting.
Respondents are then asked to report whether these activities were done for their main job or for
their “other” job. Finally, respondents are asked to report about other income-generating activities
that were not part of their job such as “selling crafts or babysitting.” Figure 3 illustrates how
summary screens on the web might be used to collect this information. The respondent would be
presented with the activities that they reported during the day. The table could highlight activity that
has already been reported as work. The respondent would then be asked to indicate if there were any
other activities that they were paid for as part of their job or another paid activity.
Figure 3.
Summary screens for work and income-generating activities
ATUS Web Design Task
4-4
ATUS Diary Design for Web Collection
4
Secondary childcare. The current ATUS interview first asks respondents about the times in the
day when children in the household are awake. Then respondents are asked “At which times during
that time period, was [CHILD] in your care?” The secondary childcare section asks this information
about three different classes of children: own household children, own non-household children, and
non-household children. Figure 4 gives an example of how the childcare summary screen can be
designed. The respondent can indicate in the final column of the table whether the child was in their
care during the day. It would also be possible to add a checkbox that allows the respondent to
indicate the child was in their care for the entire day. Additional columns can be added to the table
for additional children including own non-household children. A separate screen may be needed to
collect this information about non-household children that are not part of the roster.
Figure 4.
Summary screen for secondary childcare
Volunteering. The volunteering summary screen will be relatively straightforward. The current
ATUS interview essentially asks if the respondent did “any volunteer activities for or through an
organization.” The respondents could be presented with their reported activities and asked to
indicate which activities were done as part of volunteering.
Eldercare. Most of the eldercare questions are asked more generally about the respondent’s
provision of care to elderly persons. However, the end of the section asks did the respondent
“provide any care or assistance yesterday?” and then “At which times or during which activities did
ATUS Web Design Task
4-5
ATUS Diary Design for Web Collection
4
you provide that care or assistance yesterday?” Once again, for those who report providing elder
care, we could present a table that summarizes their activities during the day and have them indicate
during which activities they were providing care.
Probes
The ATUS interviewers have a number of follow-up probes in the diary that are used to gather more
information on specific topics and responses. One type of follow-up probe is automated; with this
probe, a pop-up screen in the CATI instrument displays the probe question text. Any activity that
lasts more than 2 hours, with the exception of sleeping, working, or attending high school, requires
the interviewer to verify the activity by asking the question: Did you stop (activity) to do anything
else at that time? Work activities that last more than 4 hours generate an instrument pop-up
requiring the interviewer to ask the respondent if he/she took any breaks of 15 minutes or longer. If
a respondent reports sleeping more than 10 hours, the interviewer must verify the duration of the
sleep activity. Another automated probe is for respondents who do not report any eating or drinking
activities in the diary. Interviewers must verify that the respondent did not eat or drink as a main
activity on the diary day. There are a number of other probes that do not automatically pop up, but
which interviewers must still ask to collect the level of detail required for activity coding. Although
we may be able to prevent some mistakes with some instructions at the beginning of the instrument,
we anticipate that some quality checks and probing will be needed in the web instrument.
There are two locations where probes could be utilized. First, the probes could be inserted in the
web diary as respondents are entering their activities throughout the day. The advantage of this is
that it trains the respondent to adhere to the best procedures for entering their activity. For example,
it would be logical to include probes about missing travel activity when the location changes
between activities. This provides information to the respondent that this level of detail is important
in the diary and may help to prevent similar errors. However, it may be best to include minimal
probing at this point of the interview since the respondent could become frustrated by multiple
prompts for corrections or additional information. This may increase the likelihood of a break-off.
Another logical place for probes is at the end of the interview on the “Activity Summary Screen”
after the respondent has had the opportunity to report all activities for the diary day. The advantage
of this location is that the probes do not interrupt the flow of the instrument for the respondent.
Some probes are only applicable to this type of screen. For example, a probe about excluding eating
or drinking can only be asked in the context of the entire day. The disadvantage of including too
many probes at the end of the interview is that the respondent may not have learned the correct
ATUS Web Design Task
4-6
ATUS Diary Design for Web Collection
4
behavior while completing the diary that leads to numerous prompts at the end of the interview for
correction. The exact placement of the probes should be guided by the importance of the probe,
whether the probe refers to the entire diary day, and how burdensome or frequent the probes might
occur and interrupt the flow of the interview.
Changes to Activity Classification
One challenging aspect of moving to web data collection is the selection of activities by respondents.
In the current ATUS interview, the respondent reports their activity verbatim and the interviewer
either uses a pre-coded category or records the response verbatim and the activity is coded later. The
role of the interviewer is helpful in this context, because they are familiar with the coding scheme.
There are different ways for respondents to select activities. One way is to have the respondent type
in the activity verbatim. This allows the respondent to describe the event in their own words, but is
the most burdensome. The task of typing is burdensome alone and verbatim recording also does not
provide the respondent much of a cue about the level of detail that the ATUS is most interested
with. Therefore, we do not recommend using verbatim recording of activities. Another way that
helps to minimize burden is to have the respondent select a pre-defined activity from a set of
dropdown boxes. Although this minimizes burden, the activity classification will need to be useable
enough for the respondent to easily find the most relevant category. In addition, these categories will
be a subset of all activity categories that may not provide the needed level of detail for ATUS. For
example, if a respondent reports reading, it might be necessary to follow-up with more specific
categories.
We create a revised activity classification by first examining the existing precodes used by
interviewers in the ATUS. These precodes correspond to activities that have only possible activity
code in the ATUS activity lexicon. Next, we analyzed the 2013 ATUS activity file to determine
which activities were most frequently reported. Table 3 shows the percentage of activities in each of
the tier 1 codes. The table lists the activities by how frequently they are reported. The “precodes”
column of the table shows how the precoded categories align with the tier 1 codes. The table shows
that almost 90 percent of the reported activities fall in the first seven categories. Hence, we focused
extensively on the types of activities that are located in these categories to make sure that
respondents could easily recognize and select the more frequent categories. Similar analyses at the
tier 2 and tier 3 level were helpful for this purpose (See Appendix D and Appendix E).
ATUS Web Design Task
4-7
ATUS Diary Design for Web Collection
4
The results of these analyses suggest a few changes to the activity classification. First, prevalent
categories such as sleeping, grooming, and watching television should be readily available at the main
activity level. Second, some categories at the second level may need to be moved within other main
activities in order to make them easier to locate for respondents. For example, “Food and Drink
Preparation” is found under “Household Activities” in the current ATUS activity classification. It
may be easier for respondents to find under “Eating and Drinking” when selecting the option on
their own.
Table 3.
Distribution of ATUS tier 1 activity codes
Tier 1
Code
Description
18
Traveling
1
Personal care
12
2
Socializing, Relaxing, and
Lesiure
Household activities
11
Eating and drinking
3
5
Caring for and helping HH
members
Work & work related activities
7
Consumer purchases
50
Precodes
Percent
Cummul.per.
19.47
19.47
Sleeping, Grooming
19.05
38.52
Watching TV
16.21
54.73
Preparing meals or snacks,
Laundry
Eating and drinking, Cleaning
kitchen
14.38
69.11
10.42
79.53
5.26
84.79
4.47
89.26
3.39
92.65
Data codes
1.28
93.93
13
Sports, exercise, and recreation
1.27
95.2
4
1.16
96.36
16
Caring for and helping non HH
members
Telephone calls
0.99
97.35
14
Religious and spiritual activities
0.91
98.26
15
Volunteer activities
0.59
98.85
6
Education
0.56
99.41
8
Professional and personal care
services
Household services
0.46
99.87
0.11
99.98
9
10
Working at main job, Working at
other job
Grocery shopping,
Attending religious services
Paying household bills
Government services & civic
obligations
ATUS Web Design Task
0.03
4-8
100
ATUS Diary Design for Web Collection
4
Third, travel is listed as a main activity. The Canadian experience demonstrates that travel is often
forgotten if it is not an explicit category. Finally, we include options for “Other” if the respondent
cannot find their activity but they want to report verbatim and “Don’t know/Can’t remember” if
they cannot find their activity or do not want to report. The suggested revision of the ATUS
classification is shown in Appendix F. As with all major revisions such as this, it will be important to
test the revised activity classification to determine its effectiveness.
In addition to analyzing the data, it would be helpful to review the literature involving ATUS data to
determine which activity categories are of analytic interest. Infrequently reported activities might be
of particular analytic interest and still worth retaining in the activity classification.
ATUS Web Design Task
4-9
This page is intentionally blank
Discussion
5
As with any major design change, such as changing modes, it will be important to evaluate different
features of the new design to both determine any differences with the existing design and to guide
the development of a more optimal design. We highlight a few areas that will need to be tested.
Contact Procedures: We have proposed a sequence of contacts that is consistent with other
projects that have attempted mixed-mode designs involving web data collection. However, we have
tailored the procedures in the context of the time use survey. For example, the timing of the
invitation and follow-ups is structured around the diary day. Experiments testing different sequences
and modes of contact would provide valuable information about the optimal timing, number, and
type of contacts. We have also proposed slightly different contact procedures for designated persons
with a high likelihood of web use versus those with a low likelihood of web use. Experiments would
lend much insight into whether the different contact procedures are helpful. It will also be
worthwhile to test the feasibility of different secure log-in procedures for the sample persons to
access the website.
Diary Design: There are many aspects of the diary design that need extensive testing and
development. First, the activity classification will be important for respondents to help them
understand the level of detail that is necessary for time use and to help them find the activities that
they participated in during the day. Cognitive testing or card sorting activities would provide a great
deal of information about the adequacy of the classification scheme.
Another area in need of testing will be the placement and implementation of probes. There will be a
trade-off between having the probes occur in the diary where they may be most relevant or on the
summary screen where they might provide the lowest level of burden throughout the interview.
Related to this feature is the implementation of how respondents can edit diary entries in a
straightforward way when mistakes are identified. Usability testing and field experiments will provide
the best evidence for placement of probes and implementation of editing procedures.
ATUS Web Design Task
5-1
Discussion
5
We have also placed the summary questions after diary entry similar to the current ATUS interview.
It may make sense to place some of these follow-ups in the diary section. For example, it might be
possible to ask extra questions during the diary after select activities such as, “Was this activity done
as part of your job or did you get paid to do this activity?” This would facilitate asking about
income-generating activities while the respondent is concentrating on a specific activity and may
eliminate the need for a summary screen. Once again, the potential improvement in measurement
would have to be balanced against the amount of burden that the extra questions in the diary place
on the respondent.
Another area that may need some testing is the respondents reaction to dependent data being
displayed in the survey interview. For example, data could be inserted into the web instrument from
the CPS regarding the respondents household or employment status. Cognitive testing or debriefing
could help to understand how respondents react to seeing this type of information.
Many of these topics will be addressed more thoroughly in the data comparability report to follow.
ATUS Web Design Task
5-2
Scope of the Mock-up Task
6
Next, Westat will create a web diary mock-up based on the design summarized in this document. As
mentioned earlier, the ATUS interview includes a number of sections collecting different types of
information. The interview includes questions about (1) the composition of the household, (2) the
respondent’s labor force and school enrollment status (3) the 24-hour time diary about “yesterday”
and (4) summary questions about specific topics of interest.
The mock-up will focus specifically on the last two components of the interview regarding the
design of the web diary and summary screens that follow. These sections of the interview collect the
most critical information for the ATUS and present the biggest challenges for moving the interview
to the web. In contrast, we believe that the first two components of the interview about household
composition and labor force status are relatively straightforward in terms of instrument design.
Hence, they are not included in the mock-up task. However, there might be elements of the diary
that interact with other parts of the interview and these will be noted in the mock-up task.
The purpose of the mock-up will be to demonstrate some important aspects of the designed diary
outlined in this task rather than providing a completely functional diary. In addition, the mock-up
will focus primarily on the design of a diary intended for administration on a laptop or desktop,
because we believe some features such as populating a summary table, may be difficult on
smartphones given the small screen size of the these devices. However, where possible, the mock-up
process will reveal potential changes that could be made in order to accommodate the smaller
devices.
ATUS Web Design Task
6-1
This page is intentionally blank
References
Bonke, J., & Fallesen, P. (2010). The impact of incentives and interview methods on response
quantity and quality in diary-and booklet-based surveys. In Survey Research Methods, 4 (2),
91-101.
Medway, R. L., & Fulton, J. (2012). When more gets you less: A meta-analysis of the effect of
concurrent web options on mail survey response rates. Public Opinion Quarterly, 76(4), 733-746.
Retrieved from http://poq.oxfordjournals.org/content/76/4/733.full.pdf+html.
Millar, M. M., and Dillman, D. A. (2011). Improving response to Web and mixed-mode surveys
Internet using address-based sampling and mail contact procedures. Public Opinion Quarterly, 75,
249-269.
.
ATUS Web Design Task
R-1
This page is intentionally blank
Part III
Comparability Report
Comparability Report
Introduction
The Bureau of Labor Statistics (BLS) is interested in adding a self-administered, web-based reporting
mode to the American Time Use Survey (ATUS) as part of a sequential mixed-mode design.
Implementing web data collection would constitute a major change in data collection procedures for
the ATUS that may impact the comparability of data from previous years with data collected with
the new design. In addition, the mixed-mode design could create discrepancies between data
collected over the web and data collected over the phone within the same year. This report identifies
some potential comparability issues along with research designs that could be used to study them.
Review of Proposed Mixed-Mode Design
Table 1 illustrates the proposed mixed-mode design from the ATUS web design task. This design
varies depending on the week that the case is in the sample, whether or not a telephone number is
available for the household, and the likelihood of the designated person being a web user. The diary
day in Table 1 refers to the actual day that the respondent reports about for the time diary.
Table 1.
Week
1
Contact procedures for mixed-mode design
2
Day
7 days before diary day
1 day after diary day
1 day after diary day
3
1 day after diary day
4-8
1 day after diary day
Comparability Report
Telephone HH
High web use
Advance mailing
Replacement letter
Replacement letter
Switch to phone for
remaining weeks;
calls begin on the day
after the diary day on
week four
Interviewer calls for
interview on day after
diary day
1
Low web use
Advance mailing
Replacement letter
Switch to phone for
remaining weeks; calls
begin on the day after
the diary day on week
three
Call for interview
Interviewer calls for
interview on day after
diary day
Non-telephone HH
Advance mailing
Replacement letter
Replacement letter
Switch to phone for
remaining weeks;
Send letter with prepaid incentive card
and call-in number
Respondent calls in
for interview on day
after diary day
Comparability Report
The proposed design includes the following features.
1. An advance mailing containing a letter and brochure describing the purpose of the ATUS.
The letter would have to be altered in order to provide the designated person with a URL,
and a user name and password or similar login procedure to use once they have reached the
ATUS diary website. It may also be helpful to provide materials that will help the respondent
complete the diary (e.g., the complete activity classification
2. We also recommend including a $2 prepaid incentive in the advanced mailing as a token of
appreciation for completing the survey. The purpose for the incentive is to encourage the
designated person to complete the interview via the web, which would lead to cost savings
versus completing the interview by telephone. Nonmonetary incentives such as a pen,
magnet, or informational brochure could also be considered if cash incentives are not
feasible; however, the literature suggests that nonmonetary incentives are not as effective as
monetary incentives (Ryu, Couper, and Marans, 2005; Singer et al., 1999, and Church, 1993).
3. A replacement letter would be sent to designated persons who did not complete the
interview during the first week.
4. Designated persons who do not complete the survey by web are eventually switched to
telephone data collection depending on their likelihood of using the web and provision of a
telephone number during the CPS.
Evaluation of Proposed Design
It is helpful to use a total survey error perspective when evaluating potential differences between
survey designs using different modes of data collection. Tourangeau, Conrad, and Couper (2013)
describe how mode can affect both errors of non-observation and errors of observation. Coverage,
nonresponse, and measurement error can all be affected by the choice of mode. A survey estimate
such as a mean consists of three components. The mode effect is simply the net effect of the
differences in errors across modes.
Survey mean = average of true values + mode effect + average of random errors (1)
Errors of non-observation consist of coverage errors due to some members of the population who
are excluded from participating with a mode (e.g., no web or telephone access) and nonresponse
errors due to those who do not respond to the survey request. Coverage errors depend on the size
of the excluded group and the size of the difference between those that are covered and those that
Comparability Report
2
Comparability Report
are not covered. Nonresponse errors depend on the size of the correlation between the likelihood of
responding in a mode and the survey variables.
Errors of observation include measurement errors, which are the difference between the survey
responses and the true values for the respondents. In a mixed-mode design the measurement errors
are a weighted average of the systematic errors in each mode.
The ideal research design for assessing mode effects will assess both errors of non-observation and
observation. Table 2 presents a research design that can estimate the effect of the proposed research
design in comparison to the current ATUS design. The proposed mixed-mode design is shown in
columns 3-5 of the table and the current ATUS design is shown in columns 6-8. The current ATUS
design has a slight modification that enables a random subsample of the interviewers to use the web
instrument when administering the interview to evaluate whether using a single instrument is
feasible. Additional training will be needed the interviewers who are administering the web
instrument. For example, the web instrument would not include analogous instructions and question
wordings that the interviewers use in the traditional CATI instrument. It is possible that the web
instrument might need to be modified slightly to assist the interviewer with administering the web
instrument.
The overall response rate from the mixed-mode design can be compared to the overall response rate
from the CATI-only design that resembles the current ATUS design to understand the effect of the
mixed-mode design on nonresponse. Relatively little is known about how response rates for web
surveys compare specifically with telephone surveys. A meta-analysis by Lozar Manfreda (2008)
included only five telephone studies, but found that response rates for web surveys are about 12
percentage points lower than their telephone counterparts. The unique design of the ATUS using
the CPS sampling frame makes it difficult to compare with other surveys.
Comparability Report
3
Comparability Report
Table 2.
Week
1
Experimental design
Day
Proposed Mixed-mode Design
Telephone
HH
High web
Low web
Non-telephone
use
use
HH
CATI-only
Telephone
HH
CATI
instrument
Advance
mailing
Replacement
letter
Advance
mailing
Replacement
letter
Advance mailing
2
1 day after diary
day
Replacement
letter
Replacement
letter
Interview calls
for interview
on day after
diary day
3
1 day after diary
day
Switch to phone
for remaining
weeks; Send
letter with prepaid calling card
and call-in
number
4-8
1 day after diary
day
Switch to
phone for
remaining
weeks; calls
begin on the
day after the
diary day on
week four
Interviewer
calls for
interview on
day after
diary day
Switch to
phone for
remaining
weeks; calls
begin on the
day after the
diary day on
week three
Interview
calls for
interview on
day after
diary day
Interviewer
calls for
interview on
day after
diary day
Respondent
calls in for
interview on day
after diary day
4
7 days before
diary day
1 day after diary
day
Replacement
letter
Advance
mailing
Interview calls
for interview
on day after
diary day
Web
instrument
Non-telephone
HH
Advance
mailing
Interview
calls for
interview on
day after
diary day
Interview
calls for
interview on
day after
diary day
Advance mailing
Interview calls
for interview
on day after
diary day
Interview
calls for
interview on
day after
diary day
Respondent
calls in for
interview on day
after diary day
Interview calls
for interview
on day after
diary day
Interview
calls for
interview on
day after
diary day
Respondent
calls in for
interview on day
after diary day
Respondent
calls in for
interview on day
after diary day
Respondent
calls in for
interview on day
after diary day
Comparability Report
We would expect the mixed-mode design would achieve higher response rates than the use of a
single telephone mode; however, some studies have found evidence for a suppression effect that
occurs when sample members refuse an initial mode such as web. For example, at least two studies
show that starting with mail and switching to web leads to higher response rates than beginning with
web and switching to mail (Messer and Dillman, 2011; Smyth, Dillman, Christian, and O’Neill,
2010). This suggests that the mail mode is the higher propensity mode and that once someone has
refused a mode such as web in a web/mail sequential mixed-mode design they are less likely to
respond by mail than if mail would have been offered first. The mode literature on sequencing of
modes is changing developing quickly and at least one other study by Olson, Smyth, and Wood
(2012) did not find this suppression effect. Furthermore, Millar and Dillman (2011) found in a study
of students that starting with web first can have an advantage when e-mail invitations are used to
contact sample members. Unfortunately, we are not aware of any relevant comparisons for web and
telephone that provide insight about the sequencing of web and telephone modes. Hence, the design
will be informative for understanding any response rate differences between the new mixed-mode
design and the current ATUS design. The design will also enable a comparison of web to telephone
response rates up until the point of the mode switch in the mixed-mode design.
Measurement error can also be evaluated by comparing estimates of time use between web and
telephone respondents up until the point of the mode switch. One potential complication with this
comparison is that any differences due to coverage and nonresponse between the modes will make it
difficult to draw conclusions about differences in measurement error. Fortunately, differences in
nonresponse and coverage across designs can be modeled using covariates from the rich CPS
sampling frame (Vannieuwenhuyze and Loosveldt, 2010). It would then be possible to use the
inverse of the response probabilities to reweight the data before conducting analyses. The
assumption behind this technique is that the respondents in the modes are statistically equivalent
conditioned on the covariates from the CPS sampling frame. This assumption would be
questionable under many survey designs, but the rich set of variables available on the CPS frame
makes the assumption more appropriate for the ATUS data.
As was discussed in the literature review and design proposal, there may be several sources of
measurement differences across modes. Some differences might point to the web having better
measurement properties and others may suggest the telephone is better. At least two factors indicate
the web has better measurement properties. First, the ATUS has been shown to have some degree
of error associated with recall (Al Bagahl et al., 2014). The diary can be completed on the web in
near real-time so that recall error is minimized. Another possibility is that the self-administered web
mode might lead to less socially desirable reporting. For example, respondents might be willing to
Comparability Report
5
Comparability Report
report more TV watching. Other factors indicate that the telephone might produce better data
quality. First, in the CATI implementation of the ATUS interview, the respondent is able to discuss
conversationally with the interviewer about their activities during the day. The interviewer may be
able to more accurately record the activities that the respondent reports than would be captured by a
categorized self-report. Second, the interviewer might be able to probe for inconsistencies better
than a web instrument, leading to better data quality. Finally, the telephone mode might lead to more
complete data since the interviewer can also probe the respondent to provided missing data. Hence,
an analysis of item nonresponse would also be important in comparing web and telephone data
collection.
Additional Experimental Factors
There are some additional experimental factors that might be considered based on the materials that
have been completed for this project and discussions between BLS and Westat. Some of these
experimental factors would involve substantial changes in the contact information available for the
ATUS.
First, a challenge with the ATUS design is the need for the respondent to complete the web diary on
a specific day. One design factor that could influence the likelihood of this occurring as planned is a
reminder on or about the day before the respondent sample member is to complete the web diary.
Hence one might consider either a reminder telephone or IVR call the day before the web diary is to
be completed. This call also provides the opportunity to give the respondent instructions about the
web diary if needed. The potential downside of this call is that it provides an opportunity for the
respondent to refuse and it also increases the cost of the mixed-mode design.
Second, there is evidence that e-mail augmentation of web requests increases the likelihood of
response via the web (Millar and Dillman, 2011). For example, the design could be altered by
sending e-mails with web links at the same time as advance mailings to increase response by the web.
This design feature would require e-mail addresses (preferably for all household members) to be
collected at the time of the CPS. It may also be possible to ask respondents to provide an e-mail
address in response to the advance mailing.
Comparability Report
6
Comparability Report
Adjustments for Mode Effects
There are different approaches that can be used to adjust the data for mode effects (Kolenikov and
Kennedy, 2014; Vannieuwenhuyze and Loosveldt, 2010; Elliot et al., 2009). However, there are very
few examples of studies where an adjustment is made for mode effects. This is probably because it
can be difficult in many situations to determine whether the adjustment procedure produces better
estimates than the unadjusted estimates. Kolenikov and Kennedy (2014) discuss some different
approaches to adjusting estimates for mode effects. They found that 19 out of 297 variables
examined showed significant differences between the CATI and web groups. One set of approaches
they used for adjustment of the estimates used regression techniques. The regression adjustment
involved regressing the survey responses on an indicator for mode and additional covariates. The
procedure then adjusts the aggregate estimate for one mode by subtracting the estimated mode
effect. Another set of approaches involved treating the responses to one mode as missing and
imputing the values based on a model. They tried two different imputation models. They found that
the regression and imputation approaches generally produced similar results in terms of adjustments
to the point estimates. Each produced a roughly four to six percent change in the estimates such
that the adjusted estimates were in the direction of less social desirability. It is reassuring that the
fairly simple regression method produces similar estimates to the relatively more computationally
complex imputation based estimates. However, Kolenikov and Kennedy argue that the larger
standard errors computed by the imputation methods more accurately reflect the uncertainty of the
adjustment.
Diary Design
There are many aspects of the diary design that will need to be tested before they are implemented
in the field. We suggest that much of this work be done in a laboratory setting using techniques such
as cognitive testing and usability testing.
First, the activity classification will be important for respondents to help them understand the level
of detail that is necessary for ATUS and to help them find the activities that they participated in
during the day. Westat proposed a revision to the ATUS activity classification in the web design task.
The revision was developed by examining the frequency of different activities and making the more
frequent activities easier to find in the classification by sometimes making them a single category.
Some activities such as food and drink preparation were moved from being under household
Comparability Report
7
Comparability Report
activities to eating and drinking to form a more logical grouping for an untrained respondent.
Cognitive testing or card sorting activities would provide a great deal of information about the
adequacy of the classification scheme. In addition, usability testing would also be helpful. For
example, observing respondents answering questions about their activities may reveal how much
detail is necessary for respondents to accurately report their activities. This would be informative for
determining how many levels of activity classification would be needed in any drop-down menus.
We also believe that a search capability should be developed for cases where the respondent cannot
find a suitable activity category. Usability testing would provide the best insights into how the search
is being used and whether the search is preferable to drop-down menus.
The placement of probes in the web instrument should also be addressed with usability testing. As
discussed in the design proposal, there will be a trade-off between having the probes occur during
the diary where they may be most relevant or on the summary screen where they might provide the
lowest level of burden throughout the interview. Usability testing would provide an understanding
of the respondents’ experience with probes in the instrument. Related to this feature is the
implementation of how respondents can edit diary entries in a straightforward way when mistakes
are identified. The ability to edit may become very complex and will have to be carefully planned
and tested. For example, changing or adding an existing entry in the web diary might create conflicts
with other entries. Usability testing will provide evidence about how respondents might prefer to
edit diary entries.
BLS staff may also conduct usability testing to explore the best place to include summary questions.
The ATUS uses summary questions after the time diary to collect more detailed information about
times when the respondent was working or volunteering. It may make sense to place follow-ups
related to income-generating activity directly in the diary section. For example, it might be possible
to ask extra questions during the diary after select activities such as, “Was this activity done as part
of your job or did you get paid to do this activity?” This would facilitate asking about incomegenerating activities while the respondent is concentrating on a specific activity and may eliminate
the need for a summary screen. The main drawback of this approach is that it adds burden to the
initial data entry task. Usability testing would provide evidence about how this placement affects the
user’s experience.
Comparability Report
8
References
Al Baghal, T., Belli, R.F., Phillips, L.A., and Butler, N.R. (2014). What are you doing now? ActivityLevel Responses and Recall Failures in the American Time Use Survey. Journal of Survey
Statistics and Methodology, 4, 519-537.
Church, A.H. (1993). Estimating the effect of incentives on mail survey response rates: A metaanalysis. Public Opinion Quarterly, 57, 62-79.
Elliott, M. N., A. M. Zaslavsky, E. Goldstein, W. Lehrman, K. Hambarsoomians, M. K. Beckett, and
L. Giordano (2009). Effects of survey mode, patient mix, and nonresponse on CAHPS®
Hospital Survey scores. Health Services Research, 44, 501–518.
Kolenikov, S., and Kennedy, C. (2014). Evaluating three approaches to statistically adjust for mode
effects. Journal of Survey Statistics and Methodology, 2, 126-158.
Lozar Manfreda, K., Bosnjak, M., Berzelak, J., Haas, I., and Vehovar, V. (2008). Web surveys versus
other survey modes: A meta-analysis comparing response rates. International Journal of Market
Research, 50, 79-104.
Messer, B. L., and Dillman, D. A. (2011). Surveying the general public over the Internet using
address-based sampling and mail contact procedures. Public Opinion Quarterly, 75, 429-457.
Messer, B. L., and Dillman, D. A. (2011). Surveying the general public over the Internet using
address-based sampling and mail contact procedures. Public Opinion Quarterly, 75, 429-457.
Millar, M. M., and Dillman, D. A. (2011). Improving response to Web and mixed-mode surveys
Internet using address-based sampling and mail contact procedures. Public Opinion Quarterly, 75,
249-269.
Olson, K., Smyth, J. D., and Wood, H. M. (2012). Does giving people their preferred survey mode
actually increase participation rates? Public Opinion Quarterly, 76, 611-635.
Ryu, E., Couper, M.P., and Marans, R.W. (2015). Survey incentives: cash vs. in-kind; face-to-face vs.
mail; response rate vs. nonresponse error. International Journal of Public Opinion Research, 18, 89106.
Singer, E., Van Hoewyk, J., Gebler, N., Raghunathan, T., & McGonagle, K. (1999). The effect of
incentives in telephone and face-to-face surveys. Public Opinion Quarterly, 64, 171-188.
Smyth, J. D., Dillman, D. A., Christian, L. M., and O’Neill, A. C. (2010). Using the Internet to
survey small towns and communities: Limitations and possibilities in the early 21st century.
American Behavioral Scientist, 53, 1423-1448.
Vannieuwenhuyze, J. T. A., Loosveldt, G., and Molenberghs, G. (2010). A method for evaluating
mode effects in mixed-mode surveys. Public Opinion Quarterly, 74, 1027-1045.
Comparability Report
R-1
This page is intentionally blank
Part IV
Final Mock-Up
Synthesis and Final Mock-up
This report discussed the potential and challenges with implementing a mixed-mode design
involving web data collection for the ATUS. We briefly synthesize the findings from the report and
introduce the final mock-up design that was created at the end of the project.
Synthesis
This literature review was undertaken to inform the design of a web instrument that could be used
in a mixed-mode design in the ATUS. We briefly reviewed the current features of the ATUS
relevant for the design. We also reviewed the literature illustrating potential sources for error related
to coverage, nonresponse, and measurement error for a web instrument in a mixed-mode design for
the ATUS. Finally, we reviewed several surveys that are using web instruments to measure time use.
The literature suggests that including web in a mixed-mode design is feasible. Other countries have
implemented a similar approach (e.g., Bonke and Fallesen, 2012); however, nearly all of the studies
we reviewed were pilot studies that have not been implemented in a production environment.
Hence, time use studies using the web relatively rare. The current literature offers some guidance
about how self-administered data collection may relate to the current telephone administered design,
but there are many areas that need further research. Further research is needed to understand if a
light diary approach with a more general activity classification produces similar results to a full diary
such as ATUS. More genereally, additional methodological research is needed to understand the best
strategies for pushing sample members to complete surveys on the web.
Beginning with the literature review and the current design of the ATUS we developed a suggested
web design for moving the ATUS to a mixed-mode design including web data collection. In this
web design task, we proposed a design includes the following features.
1.
An advance mailing containing a letter and brochure describing the purpose of the ATUS.
The letter would have to be altered in order to provide the designated person with a URL,
and a user name and password or similar login procedure to use once they have reached the
ATUS diary website. It may also be helpful to provide materials that will help the respondent
complete the diary (e.g., the complete activity classification
Comparability Report
1
Final Mock-Up
2.
3.
4.
We also recommend including a $2 prepaid incentive in the advanced mailing as a token of
appreciation for completing the survey. The purpose for the incentive is to encourage the
designated person to complete the interview via the web, which would lead to cost savings
versus completing the interview by telephone.
A replacement letter would be sent to designated persons who did not complete the
interview during the first week.
Designated persons who do not complete the survey by web are eventually switched to
telephone data collection depending on their likelihood of using the web and provision of a
telephone number during the CPS.
There are a variety of different aspects of this design that would need to be evaluated. Errors of
nonobservation (e.g. coverage and nonresponse error) and errors of observation (e.g. measurement
error) would need to be evaluated in the design. The proposed experiments in Part III of this report
facilitate the evaluation of these different error sources. The experiment compares a mixed-mode
design with the current ATUS design. The results of the experiment will be informative for
understanding differences in response through a comparison of overall response rates by condition.
Response rates up until the point of a mode switch will provide the clearest indication of differences
in response by mode. Measurement error such as differences in socially desirable reporting or item
nonresponse can also be evaluated by the design. The design also evaluates the potentional for
interviewers to use the web instrument for data collection over the phone.
We also discussed additional experimental factors to consider such as alternative contact procedures
and e-mail augmentation of contact attempts. Calling the sample member on the day before they are
to complete their web diary could increase the likelihood of completing the web diary at the
specified time. The potential downside of this call is that it provides an opportunity for the
respondent to refuse and it also increases the cost of the mixed-mode design. Second, there is
evidence that e-mail augmentation of web requests increases the likelihood of response via the web
(Millar and Dillman, 2011). The design could be altered by sending e-mails with web links at the
same time as advance mailings to increase response by the web. This design feature would require email addresses (preferably for all household members) to be collected at the time of the CPS or it
may also be possible to ask respondents to provide an e-mail address in response to the advance
mailing.
There will be additional factors related to the web design of the diary that will need to be evaluated.
One primary feature is the modification of the activity classification scheme for self administered use
on the web. Cognitive testing and card sorting activities will be helpful to understand how
Final Mock-Up
2
Final Mock-Up
respondents thinking about time use on their own. Their will be many aspects of the diary that will
help frame the task for respondents that will need to be testing including question wording and
other detailed features such as the minimum amount of time that the respondent is able to report
about. These all give cues to the respondents about the nature of the task and the level of detail in
their response. Other important features such as the placement of probes and summary questions
will be important for the new ATUS web design. Usability testing will be particularly informative
about these features.
Mock-up
The final task in this project was to create a mock-up of a potential web instrument for the ATUS.
The final mock-up is available at http://atus.wesdemo.com/. We will briefly review some of the key
features of the mock-up. The home page for the mock-up is shown below in figure 1.
Figure 1.
Home page for ATUS mock-up
The link to the PDF provides some brief instructions about how to use the mock-up and describes
some of the limitations. For example, features that are not fully functional in the mock-up are
described in the document. Pressing “Start” will begin the mock-up.
Final Mock-Up
3
Final Mock-Up
Once you begin the mock-up an overlay is shown on the screen to provide the respondent with
contextual help (see figure 2). For example, respondents can be told “If you cannot find your activity
select the closest match.”
Figure 2.
Contextual overlay
Closing the overlay enables the respondent to begin entering their activities. The mock-up features a
narrative layout that makes the time use diary more engaging for the respondent to complete. The
respondent begins by reporting their first activity. As shown in figure 3, the respondent enters the
end time of their first activity, records what activity they were doing, who they were with, and
records the location where they were at during the acivity. Time is initially shown in 5 minute
increments. The respondent can use the widget to record when they stopped the activity. The mockup also shows the duration of the activity.
Final Mock-Up
4
Final Mock-Up
Figure 3.
Recording time spent on activity
Next, the respondent records the activity that they were partipating in, who they were with, and their
location. As shown in figure 4, the respondent simply begins typing in the text boxes and as the
respondent is typing the mock-up will begin making suggestions for activities. There are only a few
activities included in the mock-up, so this functionality is somewhat limited in the mock-up.
However, the ability of the web diary to provide intelligent suggestions for activities, others who
were with the respondent, and locations would be critical to the success of the actual instrument in
production. Hence, this is a feature that would need to be enhanced and tested thoroughly in the
creation of a web instrument. The current mock-up also does not include many dependencies in the
current ATUS instrument. For example, if a respondent reports certain personal care activity the
current ATUS instrument does not ask who the respondent he or she was with or the location.
These dependencies would need to be programmed into the final ATUS web instrument.
Final Mock-Up
5
Final Mock-Up
Figure 4.
Recording activities, who was with respondent, and location
The respondent then enters all of their activities. Once the respondent is finished entering activities
and clicks done adding activities the mock-up checks for any gaps as shown in figure 5.
Figure 5.
Final Mock-Up
Checking for gaps in reported time use
6
Final Mock-Up
The mock-up then indicates the location of the gaps in the respondent’s reported time. For example,
the respondent in figure 6 did not reported activities in one time slot in the early evening and at the
end of the day. The respondent can click on these gaps to complete the time diary.
Figure 6.
Identification of gaps in reported time use
Once the respondent has completed entering all activities for the day they are taken to a summary
screen where they can review their entries for the entire day (see figure 7). This provides the
respondent with an opportunity to see their entire day in context and make any adjustments as
needed.
Final Mock-Up
7
Final Mock-Up
Figure 7.
Example summary screen
After finishing the time diary section of the instrument the mock-up shows an example of entering
secondary activity. Figure 8 shows how a respondent could indicate during which activities during
the day that a child was present by checking a box next to the activity.
Figure 8.
Final Mock-Up
Example secondary activity screen
8
References
Bonke, J., & Fallesen, P. (2010). The impact of incentives and interview methods on response
quantity and quality in diary-and booklet-based surveys. In Survey Research Methods, 4 (2),
91-101.
Millar, M. M., and Dillman, D. A. (2011). Improving response to Web and mixed-mode surveys
Internet using address-based sampling and mail contact procedures. Public Opinion Quarterly, 75,
249-269.
Comparability Report
R-1
This page is intentionally left blank
Appendix A
Paper Diary Examples
Japanese time use survey paper diary
Literature Review on ATUS Web Collection
A-1
Appendix A
Paper Diary Examples
Literature Review on ATUS Web Collection
A-2
Appendix A
Paper Diary Examples
National Statistics Omnibus Survey
Literature Review on ATUS Web Collection
A-3
Appendix A
Paper Diary Examples
Time Use in Ireland
Literature Review on ATUS Web Collection
A-4
Appendix A
Paper Diary Examples
New Zealand time use survey paper diary
Literature Review on ATUS Web Collection
A-5
Appendix A
Paper Diary Examples
Literature Review on ATUS Web Collection
A-6
Appendix A
Paper Diary Examples
Literature Review on ATUS Web Collection
A-7
Appendix A
Paper Diary Examples
Literature Review on ATUS Web Collection
A-8
Appendix A
Paper Diary Examples
Literature Review on ATUS Web Collection
A-9
Appendix A
Paper Diary Examples
Literature Review on ATUS Web Collection
A-10
Appendix A
Paper Diary Examples
Literature Review on ATUS Web Collection
A-11
Appendix A
Paper Diary Examples
Literature Review on ATUS Web Collection
A-12
Appendix A
Paper Diary Examples
Literature Review on ATUS Web Collection
A-13
Appendix A
Paper Diary Examples
This page is intentionally blank
Literature Review on ATUS Web Collection
A-14
Appendix B
Web Diary Examples
Canadian Time Use Survey Web Diary
Literature Review on ATUS Web Collection
B-1
Appendix B
Web Diary Examples
Literature Review on ATUS Web Collection
B-2
Appendix B
Web Diary Examples
Literature Review on ATUS Web Collection
B-3
Appendix B
Web Diary Examples
Literature Review on ATUS Web Collection
B-4
Appendix B
Web Diary Examples
Literature Review on ATUS Web Collection
B-5
Appendix B
Web Diary Examples
Literature Review on ATUS Web Collection
B-6
Appendix B
Web Diary Examples
Literature Review on ATUS Web Collection
B-7
Appendix B
Web Diary Examples
Literature Review on ATUS Web Collection
B-8
Appendix B
Web Diary Examples
Literature Review on ATUS Web Collection
B-9
Appendix B
Web Diary Examples
Literature Review on ATUS Web Collection
B-10
Appendix B
Web Diary Examples
Literature Review on ATUS Web Collection
B-11
Appendix B
Web Diary Examples
Literature Review on ATUS Web Collection
B-12
Appendix B
Web Diary Examples
Literature Review on ATUS Web Collection
B-13
Appendix B
Web Diary Examples
Literature Review on ATUS Web Collection
B-14
Appendix B
Web Diary Examples
Literature Review on ATUS Web Collection
B-15
Appendix B
Web Diary Examples
Literature Review on ATUS Web Collection
B-16
Appendix B
Web Diary Examples
Modular Online Time Use Survey (MOTUS)
Literature Review on ATUS Web Collection
B-17
Appendix B
Web Diary Examples
UK Millennium Study
Literature Review on ATUS Web Collection
B-18
Appendix B
Web Diary Examples
Literature Review on ATUS Web Collection
B-19
Appendix B
Web Diary Examples
Literature Review on ATUS Web Collection
B-20
Appendix B
Web Diary Examples
Literature Review on ATUS Web Collection
B-21
Appendix B
Web Diary Examples
Literature Review on ATUS Web Collection
B-22
Appendix B
Web Diary Examples
Literature Review on ATUS Web Collection
B-23
Appendix B
Web Diary Examples
Wu et al. (2012)
Literature Review on ATUS Web Collection
B-24
Appendix B
Web Diary Examples
Literature Review on ATUS Web Collection
B-25
Appendix B
Web Diary Examples
Literature Review on ATUS Web Collection
B-26
Appendix B
Web Diary Examples
Literature Review on ATUS Web Collection
B-27
Appendix B
Web Diary Examples
Literature Review on ATUS Web Collection
B-28
Appendix B
Web Diary Examples
ACT24
National Cancer Institute (NCI) has developed a physical activity evaluation tool that asks
respondents to recall and report in 5 minute increments everything they did yesterday. Respondents
must choose from a pre-populated list of activities falling into multiple categories, including
sleeping, eating, working, reading, and exercising. The combination of activity, level of excretion,
and stance/position are tied to predefined classifications of activity levels (light, moderate, vigorous)
and behavior (sleeping, sedentary, active, mixed) based on a defined scoring algorithm. F2 is the
hotkey that will end the tutorial should you want to skip.
1. Tutorial Screen 1
The tutorial pops up immediately after the respondent has entered the survey. The screen
below shows the start of the tutorial.
Literature Review on ATUS Web Collection
B-29
Appendix B
Web Diary Examples
2. Tutorial Screen 2
This screen shows the first part of the tutorial.
Literature Review on ATUS Web Collection
B-30
Appendix B
Web Diary Examples
3. Tutorial Screen 3
Tutorial screen 3 show the respondent that he or she will be reporting for a 24 hour day.
Literature Review on ATUS Web Collection
B-31
Appendix B
Web Diary Examples
4. Activity Screen 1
This screen shows the respondent entering the first activity, which is sleeping.
Literature Review on ATUS Web Collection
B-32
Appendix B
Web Diary Examples
5. Activity Screen 2
Activity screen 2 shows the respondent entering the start time and the finish time for an
activity and then choosing the type of activity and the purpose of the activity.
Literature Review on ATUS Web Collection
B-33
Appendix B
Web Diary Examples
6. Activity Screen 3
Activity screen 3 shows a gap that the respondent has created in the reporting. The red patch
is the gap.
Literature Review on ATUS Web Collection
B-34
Appendix B
Web Diary Examples
7. Prompt Screen 1 for Missing Activity
The survey has detected a gap and prompts the respondent to fill in the gap with an activity.
Literature Review on ATUS Web Collection
B-35
Appendix B
Web Diary Examples
8. Activity Screen 4 for an Activity Not on the Pick LIst
The respondent was engaged in an activity not on the pick list. The respondent typed in his
or her activity.
Literature Review on ATUS Web Collection
B-36
Appendix B
Web Diary Examples
ASA24
ASA24 is NCI’s dietary assessment tool of a 24-hour recall – either midnight to midnight yesterday
or the last 24 hours relative to when respondent starts the reporting session. The version shown in
the screen shots was programmed using Microsoft Silverlight. A new version is under development
that will be browser based and will render nicely on tablet and smartphones. The new version will
also have the option for respondents to use as a diary (i.e., a food record) where they in real time
rather than relying on recall.
1. Screen 1 is an introductory screen that explains the reporting requirements to the
respondent. The avatar is a talking penguin who explains the survey to the respondent.
Literature Review on ATUS Web Collection
B-37
Appendix B
Web Diary Examples
2. Screen shot 2 -- First Screen Where Reporting Begins
This is the first screen where the respondent can begin to enter his or her food and drink.
Literature Review on ATUS Web Collection
B-38
Appendix B
Web Diary Examples
3. Screen Shot 3 – Choose a Meal or a Snack
First the respondent enters whether the food and drink is a meal or a snack and provide
details about when and where the food or drink was consumed.
Literature Review on ATUS Web Collection
B-39
Appendix B
Web Diary Examples
4. Screen Shot 4 -- Pick List or Type in
The respondent can either pick his or her foods from a pick list or check “other” and type in
the name of the food.
Literature Review on ATUS Web Collection
B-40
Appendix B
Web Diary Examples
5. Screen Shot 5 – Foods not on the Pick List
The respondent has entered a food that is not on the pick list. This screen shows how a food
is written in.
Literature Review on ATUS Web Collection
B-41
Appendix B
Web Diary Examples
6. Screen Shot 6 – the Completed Meals for the Day
This screen shows how the foods consumed during the day are entered as either a meal or a
snack.
Literature Review on ATUS Web Collection
B-42
Appendix B
Web Diary Examples
7. Screen Shot 8 – Queries to the Respondent
The respondent is queried about eating or drinking in periods of time in which the
respondent did not report any foods or drinks.
Literature Review on ATUS Web Collection
B-43
Appendix B
Web Diary Examples
8. Screen Shot 8 –Amount for Each Food
The respondent is asked to choose an amount for reach food or drink consumed.
Literature Review on ATUS Web Collection
B-44
Appendix B
Web Diary Examples
9. Screen Shot 9 – Characteristics of the Food
The respondent is queried about the different characteristics of the food or drink he or she
reports consuming.
Literature Review on ATUS Web Collection
B-45
Appendix B
Web Diary Examples
10. Screen Shot 10 -- Further Details
The respondent is asked to report where he or she obtained the food or drink he or she
reports.
Literature Review on ATUS Web Collection
B-46
Appendix B
Web Diary Examples
11. Screen Shot 11 – Supplements
The respondent is asked to report any supplements.
Literature Review on ATUS Web Collection
B-47
This page is intentionally blank
Appendix C
Mobile Diary Examples
Dutch Time Use Survey
Literature Review on ATUS Web Collection
C-1
Appendix C
Mobile Diary Examples
NCI Life in a Day
Figure 1a. Install Screen
Literature Review on ATUS Web Collection
C-2
Appendix C
Mobile Diary Examples
Figure 1b. Install Screen (Continued)
Description: scroll down screen to see full page
Literature Review on ATUS Web Collection
C-3
Appendix C
Mobile Diary Examples
Figure 1c. Install Screen
Literature Review on ATUS Web Collection
C-4
Appendix C
Mobile Diary Examples
Figure 1d. Install Screen
Literature Review on ATUS Web Collection
C-5
Appendix C
Mobile Diary Examples
Figure 2. Application Splash Screen
Literature Review on ATUS Web Collection
C-6
Appendix C
Mobile Diary Examples
Figure 3a. Registration Screen
Description: On first open of application, the user is presented with a registration screen to capture
a user profile. First name, Last name, and Email are required fields.
Literature Review on ATUS Web Collection
C-7
Appendix C
Mobile Diary Examples
Figure 3b. Registration Screen (Continued)
Description: Scroll down to few full list. Accepting terms of service is a required field
Literature Review on ATUS Web Collection
C-8
Appendix C
Mobile Diary Examples
Figure 3c. Registration (Draft Terms of Service) Screen
Description: Placeholder screen for future terms of service language.
Literature Review on ATUS Web Collection
C-9
Appendix C
Mobile Diary Examples
Figure 4. First Time Application Launch Screen
Description: Following user registration, the user is presented with the following initial instructions
on first launch of the application
Literature Review on ATUS Web Collection
C-10
Appendix C
Mobile Diary Examples
Figure 5. Default Activities Screen
Description: Activities screen contains 23 user customizable activity buttons, one private button, and
the more activities button for logging addition activities beyond the 23 present buttons.
Literature Review on ATUS Web Collection
C-11
Appendix C
Mobile Diary Examples
Figure 6. Default Event Markers Screen
Description: The Markers screen allows the user to create and customize event buttons for logging
discrete time events and contextual data.
Literature Review on ATUS Web Collection
C-12
Appendix C
Mobile Diary Examples
Figure 7a. Daily Log Screen Zoom 1
Description: The My Log screen automatically logs activities and events for visualization, review,
and editing. The user can zoom in and out using the arrow buttons on the top left of the screen.
Literature Review on ATUS Web Collection
C-13
Appendix C
Mobile Diary Examples
Figure 7b. Daily Log Screen Zoom 2
Description: Zoom 2 provides a closer view of daily activities for reviewing or editing of short
duration activities or events
Literature Review on ATUS Web Collection
C-14
Appendix C
Mobile Diary Examples
Figure 7c. Daily Log Screen Zoom 3
Description: Zoom 3 provides a closer view of daily activities for reviewing or editing of short
duration activities or events
Literature Review on ATUS Web Collection
C-15
Appendix C
Mobile Diary Examples
Figure 8a. More Menu Screen
Description: The more menu provides an additional features, options, and information for the Life
in a Day application.
Literature Review on ATUS Web Collection
C-16
Appendix C
Mobile Diary Examples
Figure 8b. More Menu Recurring Activities Screen
Description: Recurring activities can be set during the initialization process, and can be edited at any
time using the recurring activities tab on the more menu.
Literature Review on ATUS Web Collection
C-17
Appendix C
Mobile Diary Examples
Figure 8c. More Menu My Profile Screen
Description: The users profile can be set during the initialization process, and can be edited at any
time using the My Profile tab on the more menu.
Literature Review on ATUS Web Collection
C-18
Appendix C
Mobile Diary Examples
Figure 8d. More Menu Edit Activity Buttons Screen
Description: User customized activity buttons and be edited, deleted, or reorganized using the Edit
Activity Buttons tab on the more menu.
Literature Review on ATUS Web Collection
C-19
Appendix C
Mobile Diary Examples
Figure 8e. More Menu Edit Marker Buttons Screen
Description: User customized Marker buttons and be edited, deleted, or reorganized using the Edit
Marker Buttons tab on the more menu.
Literature Review on ATUS Web Collection
C-20
Appendix C
Mobile Diary Examples
Figure 8f. More Menu Email Data Logs (Sample) Screen
Description: Life in a Day data files (currently stored in CSV format on the phones SD card) can be
emailed using the Email Data Logs tab on the more menu. Available email accounts and sharing
options well vary by phone based on user installed applications.
Literature Review on ATUS Web Collection
C-21
Appendix C
Mobile Diary Examples
Figure 8g. More Menu Delete Logs Screen
Description: Users can delete all existing data files, or selected file types, stored on the phones SD
card using the Delete Logs Screen on the more menu.
Literature Review on ATUS Web Collection
C-22
Appendix C
Mobile Diary Examples
Figure 8h. More Menu About Vignet Screen
Description: Information about the Developer of Life in a Day is provided in the About Vignet tab
on the more menu.
Literature Review on ATUS Web Collection
C-23
Appendix C
Mobile Diary Examples
Figure 9a. Customize Activity Button Workflow Screen 1
Description: To customize an activity button, the user begins from the Activities screen and taps on
any button which displays “Tap to Set”.
Literature Review on ATUS Web Collection
C-24
Appendix C
Mobile Diary Examples
Figure 9b. Customize Activity Button Workflow Screen 2
Description: The user is presented with a search screen and can type in search terms to select
specific activities.
Literature Review on ATUS Web Collection
C-25
Appendix C
Mobile Diary Examples
Figure 9c. Customize Activity Button Workflow Screen 3
Description: The search term is queried against a keywords list and matching activity descriptions
are displayed. The user can select from the list or select “I didn’t find what was looking for” and
create a new activity. After selecting an activity push the green GO button.
Figure 9d. Customize Activity Button Workflow Screen 4
Literature Review on ATUS Web Collection
C-26
Appendix C
Mobile Diary Examples
Description: Once a user selects an activity from the presented descriptions they will preview the
button with a default button label for the selected activity. The user can modify the button label by
tapping the yellow tab. This does not alter the underlying activity.
Literature Review on ATUS Web Collection
C-27
Appendix C
Mobile Diary Examples
Figure 9e. Customize Activity Button Workflow Screen 5
Description: Sample of editing feature for activity buttons.
Literature Review on ATUS Web Collection
C-28
Appendix C
Mobile Diary Examples
Figure 9f. Customize Activity Button Workflow Screen 6
Description: After the users presses the green Save tab (see figure 9d) the customized button label is
assigned to the Activities screen.
Literature Review on ATUS Web Collection
C-29
Appendix C
Mobile Diary Examples
Figure 10a. Customize Marker Button Workflow Screen 1
Description: To customize a Marker button, the user begins from the Markers screen and taps on
the button which displays “Tap to Add”, in order to add additional buttons to the screen the user
pressing the “Add Another Button” link.
Literature Review on ATUS Web Collection
C-30
Appendix C
Mobile Diary Examples
Figure10b. Customize Marker Button Workflow Screen 2
Description: Tapping on Marker Type allows the user to assign one of multiple Marker Categories
to the new button.
Literature Review on ATUS Web Collection
C-31
Appendix C
Mobile Diary Examples
Figure 10c. Customize Marker Button Workflow Screen 3
Description: The user can select the appropriate marker category for the new button from the list.
Literature Review on ATUS Web Collection
C-32
Appendix C
Mobile Diary Examples
Figure10d. Customize Marker Button Workflow Screen 4
Description: Next the user can create the new button by tapping on the Marker Name tab.
Literature Review on ATUS Web Collection
C-33
Appendix C
Mobile Diary Examples
Figure 10e. Customize Marker Button Workflow Screen 5
Description: The user types in the new marker button name.
Literature Review on ATUS Web Collection
C-34
Appendix C
Mobile Diary Examples
Figure10f. Customize Marker Button Workflow Screen 6
Description: The user can confirm and complete the new button by tapping on the green Save tab.
Literature Review on ATUS Web Collection
C-35
Appendix C
Mobile Diary Examples
Figure 10g. Customize Marker Button Workflow Screen 7
Description: New button is displayed on the screen, in order to create more button the user must
select add another button. A total of 30 marker buttons can be displayed.
Literature Review on ATUS Web Collection
C-36
Appendix C
Mobile Diary Examples
Figure11a. Activities Logging Workflow Screen 1
Description: In order to log activities using the real time diary feature, the user taps an assigned
activity button or uses the More Activities button for items not assigned to customized button.
Literature Review on ATUS Web Collection
C-37
Appendix C
Mobile Diary Examples
Figure11b. Activities Logging Workflow Screen 2
Description: Once the customized button is tapped, the stopwatch begins to log time for that
activity and captures the start time. Two activities can be selected to record concurrent activities.
Tap the timer or the activity button to stop and record an activity.
Literature Review on ATUS Web Collection
C-38
Appendix C
Mobile Diary Examples
Figure11c. Activities Logging Workflow Screen 3
Description: The user can confirm the start and end times of an event, edit times by pressing the
appropriate time tab, delete, or save the activity.
Literature Review on ATUS Web Collection
C-39
Appendix C
Mobile Diary Examples
Figure12a. Event Marker Logging Workflow Screen 1
Description: In order to log events using the real time diary feature, the user taps an assigned event
marker button.
Literature Review on ATUS Web Collection
C-40
Appendix C
Mobile Diary Examples
Figure12b. Event Marker Logging Workflow Screen 2
Description: The event is time-stamped and the assigned marker type and event name are displayed.
In addition the user can add text based notes, tag the event with current GPS coordinates, and tag
with images using the phone camera or gallery pictures.
Literature Review on ATUS Web Collection
C-41
Appendix C
Mobile Diary Examples
Figure12c. Event Marker Logging Workflow Screen 3
Description: Data entry screen for adding text based notes to an event marker.
Literature Review on ATUS Web Collection
C-42
Appendix C
Mobile Diary Examples
Figure12d. Event Marker Logging Workflow Screen 4
Description: When the user taps on Add GPS Tag the phones GPS radio will fetch current GPS coordinates.
Literature Review on ATUS Web Collection
C-43
Appendix C
Mobile Diary Examples
Figure12e. Event Marker Logging Workflow Screen 5
Description: Once a GPS tag has been added it the screen will display GPS Tag status as Added.
The user can delete the GPS tag from the event record in the database using the Delete GPS Tag
tab. When complete, the user taps the green button finished button.
Literature Review on ATUS Web Collection
C-44
Appendix C
Mobile Diary Examples
Figure13a. Daily Log Review Workflow Screen 1
Description: Users can view and edit activities and events from the daily log view, activities are
displayed as a time blocks based on start/end times and events are displayed as yellow triangles
based on the time recorded. To edit the user taps the block or triangle.
Literature Review on ATUS Web Collection
C-45
Appendix C
Mobile Diary Examples
Figure13b. Daily Log Review Workflow Screen 2
Description: Start and end times for logged activities can be modified and each activity can be
deleted or resaved.
Literature Review on ATUS Web Collection
C-46
Appendix C
Mobile Diary Examples
Figure13c. Daily Log Review Workflow Screen 3
Description: Tap the yellow triangle to view/edit an event marker.
Literature Review on ATUS Web Collection
C-47
Appendix C
Mobile Diary Examples
Figure13d. Daily Log Review Workflow Screen 4
Description: Event markers can be edited using a above screen and the user can visualize GPS
tagged locations using installed mapping functions on the phone.
Literature Review on ATUS Web Collection
C-48
Appendix C
Mobile Diary Examples
Figure14a. Add an Activity to My Log Recall Workflow Screen 1
Description: To log activities using the recall function for prior activities, the user can select the
correct day using the toggle arrows on the grey Date bar at the top of the screen. Then tap the +
icon at the top right corner of the screen.
Literature Review on ATUS Web Collection
C-49
Appendix C
Mobile Diary Examples
Figure14b. Add an Activity to My Log Recall Workflow Screen 2
Description: Next the user taps the Add a New Activity Tab.
Literature Review on ATUS Web Collection
C-50
Appendix C
Mobile Diary Examples
Figure14c. Add an Activity to My Log Recall Workflow Screen 3
Description: Next the user taps on the Activity Tab to select an activity to add to their log.
Literature Review on ATUS Web Collection
C-51
Appendix C
Mobile Diary Examples
Figure14d. Add an Activity to My Log Recall Workflow Screen 4
Description: The user can search for activities using text-based keywords. They can also tap on the
recent activities tab to view previously reported activities.
Literature Review on ATUS Web Collection
C-52
Appendix C
Mobile Diary Examples
Figure14e. Add an Activity to My Log Recall Workflow Screen 5
Description: The user selects an activity from activity descriptions matching the entered keyword
search term and taps the green go button to continue.
Literature Review on ATUS Web Collection
C-53
Appendix C
Mobile Diary Examples
Figure14f. Add an Activity to My Log Recall Workflow Screen 6
Description: The user can now modify the start and end times for the selected activity.
Literature Review on ATUS Web Collection
C-54
Appendix C
Mobile Diary Examples
Figure14g. Add an Activity to My Log Recall Workflow Screen 7
Description: Screen interface for editing activity start and end times.
Literature Review on ATUS Web Collection
C-55
Appendix C
Mobile Diary Examples
Figure 14h Add an Activity to My Log Recall Workflow Screen 8
Description: Once the activity, start time, and end times are selected the user can add text-based
notes if desired and complete by tapping on the Finished tab.
Literature Review on ATUS Web Collection
C-56
Appendix C
Mobile Diary Examples
Figure14i. Add an Activity to My Log Recall Workflow Screen 9
Description: The activity is written to the log for the selected time. If a gap of more than 2 minutes
is present in the activities listed within the log, the application highlights the time block with a pink
GAP
Literature Review on ATUS Web Collection
C-57
Appendix C
Mobile Diary Examples
Figure 15a. Add an Event to My Log Recall Workflow Screen 1
Description: To log events using the recall function for prior events, the user can select the correct
day using the toggle arrows on the grey Date bar at the top of the screen. Then tap the + icon at the
top right corner of the screen.
Literature Review on ATUS Web Collection
C-58
Appendix C
Mobile Diary Examples
Figure 15b. Add an Event to My Log Recall Workflow Screen 2
Description: The user can tap the Add a New Marker tab to add an event marker to their log.
Literature Review on ATUS Web Collection
C-59
Appendix C
Mobile Diary Examples
Figure 15c. Add an Event to My Log Recall Workflow Screen 3
Description: The user can edit all field for the new marker including the time, type, name, text notes,
and GPS and image tags. Once complete, the user taps the Finished button to save the event
marker to the log.
Literature Review on ATUS Web Collection
C-60
Appendix C
Mobile Diary Examples
Figure16a. Activities Log Gap Workflow Screen 1
Description: If a user has a gap in time within the log, they can complete their activities log by
tapping the pink time block labeled GAP.
Literature Review on ATUS Web Collection
C-61
Appendix C
Mobile Diary Examples
Figure16b. Activities Log Gap Workflow Screen 2
Description: The user taps the “Tap to Set” tab in order to select an activity they were doing for all
or part of the time GAP.
Literature Review on ATUS Web Collection
C-62
Appendix C
Mobile Diary Examples
Figure16c. Activities Log Gap Workflow Screen 3
Description: By default, the program well provide four options for activities, the previous activity in
the log, the next activity in the log, an option for I don’t remember, and other activities which will
allow the user to search for activities. Tap the appropriate tab to select.
Literature Review on ATUS Web Collection
C-63
Appendix C
Mobile Diary Examples
Figure16d. Activities Log Gap Workflow Screen 4
Description: Your selected activity will be displayed, next the user can edit the start and end times
and add text based notes if desired.
Literature Review on ATUS Web Collection
C-64
Appendix C
Mobile Diary Examples
Figure16e. Activities Log Gap Workflow Screen 5
Description: Editing screen for selecting start and end times for the activity.
Literature Review on ATUS Web Collection
C-65
Appendix C
Mobile Diary Examples
Figure16f. Activities Log Gap Workflow Screen 6
Description: Once the user selects the activity and edits or confirms the activity start and end times
they can save the activity to the log by pressing Finished.
Literature Review on ATUS Web Collection
C-66
Appendix C
Mobile Diary Examples
Figure16g. Activities Log Gap Workflow Screen 7
Description: The new activity is now displayed in the Daily Log. If more than two minutes gap
remains for reported activities the pink GAP time block well still be displayed. To fill in additional
activities simple tap the GAP time block and repeat the process.
Literature Review on ATUS Web Collection
C-67
Appendix C
Mobile Diary Examples
Description: The more activities button allows the user to log real-time activities using the
stopwatch function by selecting activities other than the 23 customizable buttons. Tap the More
Activities button to begin.
Literature Review on ATUS Web Collection
C-68
Appendix C
Mobile Diary Examples
Description: The user can search for activities using text-based keywords, or tap the grey select
recent activities tab to select from a list of previously logged activities.
Literature Review on ATUS Web Collection
C-69
Appendix C
Mobile Diary Examples
Description: Search for activities using text-based search terms and select the appropriate activity
description, or select I didn’t find what I’m looking for to create a new activity in the users database.
Literature Review on ATUS Web Collection
C-70
Appendix C
Mobile Diary Examples
Description: Once you select an activity press the green GO circle to start logging the new activity
using the stopwatch feature.
Literature Review on ATUS Web Collection
C-71
Appendix C
Mobile Diary Examples
Description: To stop recording an activity selected from the more activities button, tap the
appropriate green stopwatch timer button.
Literature Review on ATUS Web Collection
C-72
Appendix C
Mobile Diary Examples
Description: The user can confirm or edit the start and end times for the activity, add text-based
notes, delete the activity or save it to the log.
Literature Review on ATUS Web Collection
C-73
Appendix C
Mobile Diary Examples
Description: After you save the activity the stopwatch will be available to log another activity and
you can review your logged activities by tapping the My Log button at the bottom of the screen.
Literature Review on ATUS Web Collection
C-74
Appendix C
Mobile Diary Examples
Description: Newly logged activities are displayed in the log view for review or editing.
Literature Review on ATUS Web Collection
C-75
This page is intentionally blank
Appendix D
Most Prevalent ATUS Activity Codes
Using the Second Level of ATUS Activity Codes
Tier2 code
1203
Description
Precodes
Percent
Cummul.per.
Watching TV
12.9
12.9
101
Socializing relaxing and leisure: Relaxing
and leisure
Personal care: Sleeping
Sleeping
11.41
24.31
1101
Eating and drinking: Eating and drinking
10.37
34.68
102
Personal care: Grooming
Eating and
drinking
Grooming
7.02
41.7
202
Household activities: Food and drink prep.
6.26
47.96
1807
4.98
52.94
4.52
57.46
501
Travel: Travel related to consumer
purchases
Care for HH members: Care for HH
children
Work: Working
4.26
61.72
201
Household activities: Housework
3.39
65.11
701
Consumer purchases: Shopping
3.39
68.5
1201
Socializing relaxing and leisure: Socializing
and communicating
Travel: Travel related to work
2.93
71.43
2.91
74.34
2.8
77.14
1.98
79.12
1.92
81.04
1.89
82.93
206
Travel: Travel related to socializing,
relaxing, and leisure
Household activities: Household
management
Travel: Travel related to eating and
drinking
Travel: Travel related to care for HH
members
Household activities: Animals and pets
1.47
84.4
5001
Data codes: unable to code
1.28
85.68
1301
Sports, exercise, and recreation:
participating in sports
Travel: Travel related to care for nonHH
members
Telephone calls: Telephone calls
1.19
86.87
1.06
87.93
0.99
88.92
0.91
89.83
0.76
90.59
301
1805
1812
209
1811
1803
1804
1601
1401
1802
Working at main
job, Working at
other job
Cleaning kitchen,
laundry
Grocery shopping
Religious and spiritual activities:
Religious/spiritual practices
Travel: Travel related to household
activities
ATUS Web Design Task
Preparing meals
or snacks
Paying household
bills
Attending
religious service
D-1
This page is intentionally blank
Appendix E
Most Prevalent ATUS Activity Codes
Using the Third Level of ATUS Activity Codes
Percent
Cummul.per.
10101
Tier3 code
Personal care: Sleeping: Sleeping
Sleeping
11.16
11.16
110101
Eating and drinking: Eating and drinking:
Eating and drinking
Socializing, relaxing, and leisure: Relaxing
and leisure: Television and movies (not
religious)
Personal care: Grooming: Washing,
dressing, and grooming oneself
Household activities: Food and drink prep.:
Food and drink prep.
Work: Work: Work, main job
Eating and
Drinking
Watching TV
10.37
21.53
7.98
29.51
Grooming
7.02
36.53
Preparing meals or
snacks
Working at main
job
4.66
41.19
4.08
45.27
2.93
48.2
2.84
51.04
2.29
53.33
2.27
55.6
1.92
57.52
1.69
59.21
1.69
60.9
1.67
62.57
1.61
64.18
1.6
65.78
1.49
67.27
1.49
68.76
1.3
70.06
1.25
71.31
120303
10201
20201
50101
120101
180501
180704
30101
181101
20101
181201
120312
70104
180301
20203
180701
120301
20102
Description
Precode
Socializing, relaxing, and leisure:
Socializing and communicating:
Socializing and communicating with
others
Travel: Travel related to work: Travel
related to work
Travel: Travel related to consumer
purchases: Travel related to shopping
Care for HH members: Care for children:
Physical care for HH children
Traveling: Travel related to eating and
drinking: Travel related to eating and
drinking
Household activities: Housework: Interior
cleaning
Travel: Travel related to socializing,
relaxing, and leisure: Travel related to
socializing and communicating
Socializing, relaxing, and leisure: Relaxing
and leisure: Reading for personal interest
Consumer purchases: Shopping:
Shopping, except groceries food and gas
Traveling: Travel related to care for HH
members: Travel related to caring for HH
children
Household activities: Food and drink prep.:
Kitchen and food clean-up
Traveling: Travel related to consumer
purchases: Travel related to grocery
shopping
Socializing, relaxing, and leisure: Relaxing
and leisure: Relaxing, thinking
Household activities: Housework: Laundry
ATUS Web Design Task
E-1
Cleaning kitchen
Laundry
Appendix E
Most Prevalent ATUS Activity Codes
Using the Third Level of ATUS Activity Codes
Tier3 code
20902
20601
70101
180703
120308
30112
500101
180405
181301
70103
120307
181401
20501
10301
181203
30103
20602
140101
160101
Description
Precode
Household activities: Household
management: Household & personal
organization and planning
Household activities: Animals and pets:
Care for animals and pets
Consumer purchases: Shopping: Grocery
shopping
Traveling: Travel related to consumer
purchases: Travel related to purchasing
food
Socializing, relaxing, and leisure: Relaxing
and leisure: Computer use for leisure
Care for HH members: Care for children:
Picking up/dropping off children
Data codes: Unable to code: Insufficient
detail in verbatim
Traveling: Travel related to care for non HH
members: Travel related to helping non
HH adults
Traveling: Travel related to sports,
exercise, and recreation: Travel related to
participating in sports/exercise/rec.
Consumer purchases: Shopping:
Purchasing food (not groceries)
Socializing, relaxing, and leisure: Relaxing
and leisure: Playing games
Traveling: Travel Related to
Religious/Spiritual Activities: Travel
related to religious/spiritual practices
Household Activities: Lawn, Garden, and
Houseplants: Lawn, garden, and
houseplant care
Personal Care: Health-related Self Care:
Health-related self care
Traveling: Travel Related to Socializing,
Relaxing, and Leisure: Travel related to
relaxing and leisure
Caring For & Helping Household (HH)
Members: Caring For & Helping HH
Children: Playing with HH children, not
sports
Household Activities: Animals and Pets:
Walking / exercising / playing with
animals
Religious and Spiritual Activities:
Religious/Spiritual Practices: Attending
religious services
Telephone Calls: Telephone Calls (to or
from):Telephone calls to/from family
members
ATUS Web Design Task
E-2
Grocery shopping
Percent
Cummul.per.
1.04
72.35
0.96
73.31
0.89
74.2
0.88
75.08
0.84
75.92
0.8
76.72
0.73
77.45
0.7
78.15
0.66
78.81
0.65
79.46
0.64
80.1
0.63
80.73
0.62
81.35
0.6
81.95
0.55
82.5
0.53
83.03
0.51
83.54
0.46
84
0.46
84.46
Appendix E
Most Prevalent ATUS Activity Codes
Using the Third Level of ATUS Activity Codes
Tier3 code
180209
20104
20904
189999
500106
140102
181501
40507
60301
130131
180401
180702
20903
160102
181204
10102
30201
Description
Precode
Percent
Cummul.per.
Traveling: Travel Related to Household
Activities: Travel related to household
management
Household Activities: Housework: Storing
interior HH items, Inc. food
Household Activities: Household
Management: HH & personal e-mail and
messages
Traveling: Traveling, n.e.c.:Traveling, n.e.c.
0.43
84.89
0.37
85.26
0.37
85.63
0.36
85.99
Data Codes: Unable to Code: Gap/can’t
remember
Religious and Spiritual Activities:
Religious/Spiritual Practices: Participation
in religious practices
Traveling: Travel Related to Volunteer
Activities: Travel related to volunteering
Caring For & Helping Nonhousehold
(NonHH) Members: Helping Nonhousehold
Adults: Picking up/dropping off nonhh
adult
Education: Research/Homework:
Research/homework for class for degree,
certification, or licensure
Sports, Exercise, and Recreation:
Participating in Sports, Exercise, or
Recreation: Walking
Traveling: Travel Related to Caring For &
Helping Nonhh Members: Travel related to
caring for and helping nonhh children
Traveling: Travel Related to Consumer
Purchases: Travel related to purchasing
gas
Household Activities: Household
Management: HH & personal mail &
messages (except e-mail)
Telephone Calls: Telephone Calls (to or
from):Telephone calls to/from friends,
neighbors, or acquaintances
Traveling: Travel Related to Socializing,
Relaxing, and Leisure: Travel related to
arts and entertainment
Personal Care:Sleeping:Sleeplessness
0.36
86.35
0.35
86.7
0.35
87.05
0.34
87.39
0.32
87.71
0.32
88.03
0.32
88.35
0.32
88.67
0.31
88.98
0.27
89.25
0.26
89.51
0.25
89.76
Caring For & Helping Household (HH)
Members: Activities Related to HH
Children’s Education: Homework (hh
children)
0.24
90
ATUS Web Design Task
E-3
This page is intentionally blank
Appendix F
Revised ATUS Activity Classification
Sleeping
Washing, dressing, and grooming oneself
Watching TV
Eating and Drinking
–
Eating or Drinking – meals, snacks, drinks
–
Food and drink preparation
–
Kitchen and food clean-up
Socializing, Relaxing, and Leisure
–
Socializing and communicating with others
–
Reading for personal interest
–
Relaxing/thinking
–
Computer use for leisure
–
Playing games
–
Arts and entertainment
Personal Care
–
Grooming
–
Health-related self-care
–
Personal, private activities
–
Personal care emergencies
ATUS Web Design Task
E-1
Appendix F
Revised ATUS Activity Classification
Work
–
Working at main job
–
Working at other jobs
–
Other income generating activity
–
Job search and interviewing
Household Activities
–
Interior cleaning
–
Laundry
–
Sewing
–
Indoor or outdoor maintenance, repair, and decoration
–
Lawn, garden, and houseplants
–
Animals and pets
–
Vehicles
–
Appliances, tools, and toys
–
Household management
Shopping
–
Grocery shopping
–
Purchasing gas
–
Purchasing food (not groceries)
–
Other shopping
Travel
Final Mock-Up
F-2
Appendix F
Revised ATUS Activity Classification
Care for Household Members
–
Use existing ATUS tier 2 codes
o
Care for non-Household Members
–
Use existing ATUS tier 2 codes
o
Use existing ATUS tier 2 codes
o
Use existing ATUS tier 2 codes
o
Use existing ATUS tier 2 codes
o
Use existing ATUS tier 2 codes
o
Use existing ATUS tier 3 codes
Government and Civic Obligations
–
Use existing ATUS tier 3 codes
Household Services
–
Use existing ATUS tier 3 codes
Professional and Personal Care Services
–
Use existing ATUS tier 3 codes
Education
–
Use existing ATUS tier 3 codes
Use existing ATUS tier 3 codes
Sports, Exercise, and Recreation
–
Participating in sports, exercise, or recreation
–
Attending sporting or recreational events
Religious and Spiritual Activities
–
Use existing ATUS tier 2 codes
o
Final Mock-Up
Use existing ATUS tier 3 codes
F-3
Appendix F
Revised ATUS Activity Classification
Volunteer Activities
–
Use existing ATUS tier 2 codes
o
Use existing ATUS tier 3 codes
Telephone Calls
Other
Don’t know/ Can’t Remember
Final Mock-Up
F-4
| File Type | application/pdf |
| Author | Pat Nolan |
| File Modified | 2015-07-17 |
| File Created | 2015-07-17 |