README: working-from-home dataset

This is a README file for the working-from-home dataset.
Used in case study 20B

Data source

The data was collected as part of the experiment run by Nicholas Bloom, James Liang, John Roberts, and Zhichun Jenny Ying, published in the Quarterly Journal of Economics, “Does working from home work? Evidence from a Chinese experiment.” Volume 130, Issue 1, February 2015, Pages 165–218 HERE

The data is available HERE wfh.zip as the “Data” link to the entry “Does working form home work?”


You can use this dataset for educational purposes.


Raw data tables

There are many raw data tables. Of them, we used the following:

quit_data observations: person-level (n=249) ID variable personid other important variables expgroup whether in treatment group quitjob whether quit job by end of experiment perform11 performance z-score during experiment perform01 performance z-score before the experiment demographic variables (age, gender, whether married, whether has children)

tc_comparison observations: person-level (n=249) ID variable personid other variables all background variables used to check balance (except pre-experiment performance z-score)

performance_during_exper.dta observations: person x week level n=112,279, 1,934 persons, 86 weeks, unbalanced ID variables personid, year_week other important variables expgroup whether in treatment group (missing if not inexperiment) experiment_treatment whether in treatment group and week during exierment experiment_control whether in control group and week during exierment phonecallraw number of phone calls placed (if order taker)

Tidy data table

wfh_tidy_person a person-lvel file with people in the experiment, n=249 ID variable personid other important variables treatment whether in treatment group ordertaker whether in treatment group quitjob whether quit job by end of experiment phonecalls1 numer of phone calls placed during experiment