Hello helpful list,
I have a sample of companies where each company was treated in a single year sometime between 1990-2010 (i.e., staggered treatments). The treatment was a random audit where the firm was selected randomly by an authority at any given year.
I want to test the effect of the treatment on the treated companies over a 4 year observation window (change from 2 years before the treatment to 2 years after). But I do not want to use a matched sample. I have 200 treated companies and 3000 control companies (all other companies that were not treated).
For the treated companies, I have a variable -- time_period -- that indicates the number of years before or after the treatment (e.g., value = -2 if the calendar year is two years before the treated firm was audited).
My question: How do I create the same time_period variable for the control firms? I am confused because the control firms do not have a treatment date. Also, if I do not create the time_period variable for the control firms then how do I test a 4-year observation window? Do I always include all control firm years in the sample regardless of the observation window for the treated firms?
Thank you in advance!
Roger
I have a sample of companies where each company was treated in a single year sometime between 1990-2010 (i.e., staggered treatments). The treatment was a random audit where the firm was selected randomly by an authority at any given year.
I want to test the effect of the treatment on the treated companies over a 4 year observation window (change from 2 years before the treatment to 2 years after). But I do not want to use a matched sample. I have 200 treated companies and 3000 control companies (all other companies that were not treated).
For the treated companies, I have a variable -- time_period -- that indicates the number of years before or after the treatment (e.g., value = -2 if the calendar year is two years before the treated firm was audited).
My question: How do I create the same time_period variable for the control firms? I am confused because the control firms do not have a treatment date. Also, if I do not create the time_period variable for the control firms then how do I test a 4-year observation window? Do I always include all control firm years in the sample regardless of the observation window for the treated firms?
Thank you in advance!
Roger
Comment