No announcement yet.
  • Filter
  • Time
  • Show
Clear All
new posts

  • Difference-in-difference on matched data with binary outcome variable

    Dear Statalisters,

    I have a balanced data with three waves. I want to measure the impact of participation on a binary outcome variable. I match the pooled of the three waves based on the decision to participate. I understand how to go about applying the diff-in-diff estimator on the matched data using the -xtreg, fe- routine in Stata. This would account for the influence of time invariant unobservables in addition to the matching on observables. The corresponding routine -xtprobit- for a binary outcome does not seem to have a fixed effects option (-,fe-) equivalent. How can I apply the diff-in-diff estimator when the outcome variable is binary?

    Help is very much appreciated!