This Stata code uses the trip tour methodology described in "Measuring Commuting in the American Time Use Survey" to construct a measure of commuting from ATUS activity-level files. If you incorporate or adapt this methodology into your analysis, please cite the latest version of this paper, published in 2019 in the Journal of Economic and Social Measurement, vol. 44, no. 1, pp. 1-17. An ungated version of the paper is available at https://ideas.repec.org/p/pra/mprapa/93239.html.
atus_commuting_calculation.do calls each of the component programs in order.
The provided paring program reduces the sample in certain ways that were needed for the paper but may not be needed for your analysis. In particular, age and weekday filters may not be desired.
ATUS data were sourced from ATUS-X, in the form of hierarchical extracts. The provided .do files were edited to construct from these hierarchical files a rectangular dataset, with individual and household characteristics attached to each activity line in a respondent's time diary.
Users wishing to use their own ATUS-X input files should place the .dta files in the data/input directory and modify the accompanying .do files from ATUS-X using the guidance in the comments of the data read-in program.
A data file is provided in .zip format in the data/input directory. To use this file, unzip it in its given directory.
If you wish to incorporate other covariates, the simplest solution is to obtain a person-level rectangular extract in Stata format from ATUS-X with the desired characteristics. These characteristics can be merged using the caseid variable, which uniquely identifies households in the ATUS-X data:
merge 1:1 caseid using ../data/output/atus_commutes