Linking ATUS and CPS Files

The Relationship between ATUS and CPS

All American Time Use Survey (ATUS) respondents were previously in the Current Population Survey (CPS). ATUS data are collected throughout the year, but each ATUS respondent only completes one time diary. All time diaries collected in a given year are included in a single data file corresponding to the year that the interview was conducted. In addition to including information about what people did over a 24-hour period, the ATUS data also include information about the month of ATUS participation and the month and year that they completed the CPS. Given the regularity of the CPS rotation pattern, one can use the information available in the CPS to determine when ATUS respondents should have previously appeared in the CPS.

The CPS is a short panel survey which focuses on labor force participation but also collects information on a variety of supplemental topics such as volunteering, food security, and tobacco use. CPS respondents participate in the survey up to eight times over a 16-month period (see IPUMS CPS documentation). Each CPS observation is indexed by a "month-in-sample" or MISH that begins at one and increments by one each month a CPS household should be interviewed. After the eighth CPS interview, a subsample of CPS respondents is selected to participate in the ATUS. Participation in ATUS occurs in the two to five months following completion of the CPS.

Linking the ATUS to the CPS opens up novel opportunities for research. For instance, researchers could examine short-term life course transitions -- like finishing high school, losing a spouse to death, or exiting paid work -- over the course of the CPS. There are also opportunities to combine information from CPS topical supplements with ATUS data on time use.

Back to Top

Resources for Linking

Linkage between the ATUS and CPS, however, requires understanding the CPS rotation pattern and the varying two to five month gap between CPS and ATUS participation as well as combining data from the ATUS and CPS. We have the following resources available to help users link the CPS and ATUS.

1. RoPES

RoPES is a tool we built to help researchers understand the systematic way that CPS respondents are surveyed eight times over a 16-month period. Understanding the CPS rotation pattern is very important for understanding when ATUS respondents were previously in the CPS and the months where one can expect to observe them in the ATUS. We also provide this instructional video describing the CPS rotation pattern and use of RoPES:

2. ATUS-CPS Linking Webinar

We previously held a webinar on linking CPS observations to one another and linking CPS to ATUS. Researchers who are already familiar with the CPS rotation pattern and want to focus solely on the CPS to ATUS linkage could skip forward to the 35-minute mark

3. ATUS-CPS Linking Counts Tool

ATUS-CPS Linking Counts is a tool we built to help researchers understand how many ATUS respondents from each year previously appeared in each month and year of the CPS. Given that ATUS data are collected throughout the year and that there is a two to five month gap between the CPS and the ATUS, exactly when ATUS respondents finished the CPS varies within any given year of ATUS data. To further complicate matters, the CPS rotation pattern means that we can approximate the number of ATUS respondents who previously appeared in any month of the CPS.

The linking tool displays the number of ATUS respondents who should link to each CPS month given three different linkage windows.

1. 2-5 months prior to ATUS (final month [MIS 8] of CPS) (default)
2. 2-8 months prior to ATUS last four months [MIS 5-8] of CPS)
3. 2-20 months prior to ATUS (any month [MIS 1-8] of CPS)

Users should note that with the shortest linking window -- 2 to 5 months prior to the ATUS -- the counts displayed are the actual number of ATUS respondents who appear in CPS months. Under the second and third linking windows, the counts displayed are the number of ATUS respondents that one could expect to observe in each CPS month. Note that the number of actual linkages will likely be lower for two reasons. First, non-response is an issue; an entire household simply may not respond to the CPS in a one month and then re-enter the survey later in the CPS rotation pattern. Second, migration is also a possibility. The CPS sampling unit is the physical dwelling. Sometimes individuals in the household change as household members move in or out. Other times entire households turn over.

The counts we display for the second and third linkage windows represent the maximum number of ATUS to CPS linkages for a given ATUS year to a specific CPS month. The counts are not adjusted for eligibility or participation in ATUS modules or CPS supplements and they do not account for non-response or migration.

4. CPSIDP Linking Key

CPSIDP is a linking key that we create at IPUMS and add to both the CPS and ATUS data. CPSIDP may be used to link the ATUS to CPS. To use this variable, researchers must request data separately from both the ATUS and CPS data access systems and be sure to include CPSIDP in both of the datasets you request! When performing the merge between the two datasets, remember that each ATUS respondent can appear in the CPS up to eight times.

Back to Top

Considerations for Weighting

Weights require consideration when linking ATUS and CPS data. The ATUS weights provide representative estimates of annual time use for ATUS respondents.

Linking ATUS Respondents to their Final CPS Interview, 2-5 months prior to ATUS

When linking ATUS respondents to their final CPS interview, ATUS weights are appropriate. By definition, each ATUS respondent previously participated in a final CPS interview.

Linking ATUS Respondents to a Specific CPS Supplement

When linking ATUS respondents to a specific CPS supplement, for example, food security, ATUS weights should be adjusted. There are four reasons for this. First, not all ATUS respondents from a given year will appear in the supplement of interest. Second, for ATUS respondents who do appear in the supplement of interest, there are different time gaps between CPS and ATUS participation. Third, not every individual may have been eligible to participate in the supplement. While this is not an issue for the food security supplement, it is an issue for other supplements like fertility and job tenure, which are only asked of eligible CPS respondents. Fourth, even some individuals who were eligible to participate in CPS supplements may have declined to participate.

Weight adjustments will help account for possible differences in the characteristics of those in the linked data set for whom supplement data are available versus those for whom no usable supplement data were obtained.

Back to Top

Pre-Linked ATUS-CPS Supplement Files

A small set of data files containing CPS supplement records to which ATUS identifiers have been added may be downloaded from this page. These files are intended to help users ensure that their code is producing the same values as the procedures we describe, which are appropriate for analyses that use CPS supplement data together with time use variables.

• General documentation describing the linked ATUS-CPS supplement files
• Sample code for merging ATUS and CPS supplement files
  • SAS
  • Stata
  • SPSS

In the table below, click on:

• "Linked file" to obtain the linked ATUS-CPS supplement data
• "Replicate weights" to obtain additional weight data needed if you plan to use the survey procedures included in your preferred statistical package
• "Original Supplement Documentation" to access supplement documentation provided by the Census Bureau.
• "Linked File Case Counts" to view descriptive tabulations of the linked file.

Supplement Topic:

• Education
• Food Security
• Volunteer

School Enrollment
Supplement Month and Year	Data Files	Syntax Files			Original Supplement Documentation	Linked File Case Counts	Combine with ATUS Samples
October 2002	linked file	SAS	SPSS	STATA	pdf	pdf	2003-2004
	replicate weights	SAS	SPSS	STATA
October 2003	linked file	SAS	SPSS	STATA	pdf	pdf	2003-2005
	replicate weights	SAS	SPSS	STATA
October 2004	linked file	SAS	SPSS	STATA	pdf	pdf	2004-2006
	replicate weights	SAS	SPSS	STATA
October 2005	linked file	SAS	SPSS	STATA	pdf	pdf	2005-2007
	replicate weights	SAS	SPSS	STATA
October 2006	linked file	SAS	SPSS	STATA	pdf	pdf	2006-2008
	replicate weights	SAS	SPSS	STATA
October 2007	linked file	SAS	SPSS	STATA	pdf	pdf	2007-2009
	replicate weights	SAS	SPSS	STATA

Food Security
Supplement Month and Year	Data Files	Syntax Files			Original Supplement Documentation	Linked File Case Counts	Combine with ATUS Samples
December 2001	linked file	SAS	SPSS	STATA	pdf	pdf	2002-2003
	replicate weights	SAS	SPSS	STATA
December 2002	linked file	SAS	SPSS	STATA	pdf	pdf	2003-2004
	replicate weights	SAS	SPSS	STATA
December 2003	linked file	SAS	SPSS	STATA	pdf	pdf	2004-2005
	replicate weights	SAS	SPSS	STATA
December 2004	linked file	SAS	SPSS	STATA	pdf	pdf	2005-2006
	replicate weights	SAS	SPSS	STATA
December 2005	linked file	SAS	SPSS	STATA	pdf	pdf	2006-2007
	replicate weights	SAS	SPSS	STATA
December 2006	linked file	SAS	SPSS	STATA	pdf	pdf	2007-2008
	replicate weights	SAS	SPSS	STATA
December 2007	linked file	SAS	SPSS	STATA	pdf	pdf	2008-2009
	replicate weights	SAS	SPSS	STATA

Volunteer
Supplement Month and Year	Data Files	Syntax Files			Original Supplement Documentation	Linked File Case Counts	Combine with ATUS Samples
September 2002	linked file	SAS	SPSS	STATA	pdf	pdf	2003-2004
	replicate weights	SAS	SPSS	STATA
September 2003	linked file	SAS	SPSS	STATA	pdf	pdf	2003-2005
	replicate weights	SAS	SPSS	STATA
September 2004	linked file	SAS	SPSS	STATA	pdf	pdf	2004-2006
	replicate weights	SAS	SPSS	STATA
September 2005	linked file	SAS	SPSS	STATA	pdf	pdf	2005-2007
	replicate weights	SAS	SPSS	STATA
September 2006	linked file	SAS	SPSS	STATA	pdf	pdf	2006-2008
	replicate weights	SAS	SPSS	STATA
September 2007	linked file	SAS	SPSS	STATA	pdf	pdf	2007-2009
	replicate weights	SAS	SPSS	STATA

Back to Top