Hello!
I have an issue with some data I want to work with. I like to use panel data to execute a diff-in-diff analysis. I have iOS app-store data for 6 periods (each 6 month apart). I want to analyse, how many unaffected apps a developer supplies before and after the app function got incorperated into the iOS as an in-iOS feature (e.g. flashlight apps or QR-Scan apps). Therefore, the treatment is the introduction of the feature into the OS (without downloading an extra app). The unit of analysis is at the developer-period level. My treatment-group are all developers that have at least one app which is affected, my control-group are all developers which have no affected apps. My treamtent indicator is 1 if the app got released after the iOS update which introduced this function. My problem is, that I have 2 observations for every developer-period unit. One with the amount of apps before treatment (with the treatment indicator variable equals 0) and one with the amount of apps after treatment.
The issue is, that I can´t run a panel data regression with 2 individual-time specific units each observation (one before and after treatment).
Can someone help me how to handle data where I got 2 different kinds of time units (period I observed the app store and release date of the app)?
It is my first time ill do my own analysis outside a specific task in class and hope this question makes sense.
Thank you and all the best
Vincent