Technical information on the impact of the coronavirus (COVID-19) pandemic on the ATUS for users of ATUS 2020 public use data files
The 2020 American Time Use Survey (ATUS) was greatly affected by the coronavirus (COVID-19) pandemic. The following technical information about the public use microdata files and how they were impacted by the pandemic is intended for researchers who plan to conduct their own analyses of these data. This information is relevant to both the 2020 annual and the multiyear public use data files.
1. How did the suspension of data collection impact ATUS 2020 public use microdata files?
Data represent only 10 months of 2020. There are no data available for the period March 18, 2020 to May 9, 2020 because data were not collected about time use on these days. Researchers can use the variable TUDIARYDATE to identify the date about which respondents reported their use of time. The data files contain time diary data collected before and after the temporary suspension of ATUS data collection, but there are no values of TUDIARYDATE for March 18, 2020 to May 9, 2020.
2. Were any changes made to the ATUS weighting methodology in 2020?
Yes. Time-use estimates produced with the 2020 statistical weights represent only the days before and after the pandemic-related suspension of data collection. That is, they represent the periods January 1st to March 17th and May 10th to December 31st. In prior years, ATUS statistical weights represented all days between January 1st and December 31st.
ATUS weights are designed to produce estimates of average time use that are representative of the U.S. civilian noninstitutional population age 15 and over and various subpopulations. In prior years, ATUS statistical weights were constructed to represent the number of person-days per quarter, where a person-day refers to the number of people in a population and the number of days in the quarter. When the weights for all respondents in a given quarter are summed, the result is equivalent to the number of people in the total population times the number of days in the quarter. The weights sum to the number of person-days per quarter for the general population and selected subpopulations. Weights for weekdays and weekend days sum to the number of person-days on weekdays and weekend days for these same populations. This method was used to create weights for the third and fourth quarters of 2020.
However, the ATUS weighting methodology for the 1st and 2nd quarters was modified because of the suspension of ATUS data collection. Rather than representing all days in the first and second quarters of 2020, weights were constructed to represent only those days, weekend days, and weekdays about which the time-use data were collected. In the first quarter of 2020, weights were constructed to represent the number of person-days in the period January 1st through March 17th only. In the second quarter of 2020, weights were constructed to represent the number of person-days in the period May 10th through June 30th only. On the 2020 annual and multiyear public use microdata files, the 2020 weighting variable is called TU20FWGT. To facilitate analyses with 2019 data, this weighting variable also was created for the 2019 ATUS data. For more information about ATUS statistical weights, see chapter 7 of the ATUS User's Guide.
3. What estimates can be produced with the 2020 ATUS public use files?
While it is impossible to produce annual estimates with the 2020 ATUS data, researchers can produce partial-year estimates using data collected before or after the pandemic-related suspension of data collection. Chapter 7 of the ATUS User's Guide provides formulas for generating estimates with the ATUS data. When using the 2020 ATUS data, researchers should use the weighting variable TU20FWGT instead of TUFINLWGT.
Note that the formulas for producing estimates of the number of persons in a population and the number of participants in an activity require researchers to account for the number of days in the period. Instead of representing all 366 days in 2020, the data represent 313 days total. This amounts to 77 days in the first quarter (including February 29th), 52 days in the second quarter, 92 days in the third quarter, and 92 days in the fourth quarter.
4. Are there any new variables on the ATUS public use microdata files?
While no questions were added to the ATUS in 2020, the Bureau of Labor Statistics added questions to the Current Population Survey (CPS) to help gauge the effects of the pandemic on the labor market beginning in May 2020. Because ATUS respondents are members of households that had completed CPS, the ATUS-CPS file includes information from these new questions.
These questions ask whether people teleworked or worked at home because of the pandemic; whether people were unable to work because their employers closed or lost business due to the pandemic and, if so, whether they were paid for this missed work; whether the pandemic prevented job-seeking activities; and whether they needed medical care for something other than the coronavirus, but did not get this care because of the pandemic. All of these supplemental questions follow the CPS monthly labor force questions and refer to the 4 weeks prior the interview.
The variables PTCOVID1, PTCOVID2, PTCOVID3, PTCOVID4, and PTCOVID5W were added to the ATUS-CPS file for persons in households that were asked these questions. Cases whose 8th CPS interview occurred before May 2020 will have missing values for these variables on the ATUS-CPS file. For more information about these and other variables collected in the CPS, see the 2020 ATUS-CPS data dictionary.
Last Modified Date: July 22, 2021