Hi Everyone,
I have a dataset where there are individuals and there are multiple projects for each individual and I have an external treatment (shock or spam) that has impacted certain projects. So each project has a product owner and multiple team members.
I was wondering how I can use a DID regression in this case. I can create a variable for the product owners who got impacted and pre and post. However, the owners have multiple projects and they may not have been impacted in all the projects. So in the dataset you can see that the external shock occurred in only one product which is apple 12 and hence the owner of the projects is labelled to subject to treatment, however, she/he is not subjected to treatment in the android4 product and hence not sure how to set this as a DID experiment
dataset id product_id owner_treatment post external_shock owner_performance 1 apple12 1 0 1 34 1 android4 1 0 0 54 1 apple12 1 1 1 32 1 android4 1 1 0 21 2 faceb34 0 0 0 345 2 faceb34 0 1 0 454 Any suggestions how I can set up this as DID experiment? And how to do parallel trends? Because there are multiple projects for each individual. So not sure how to set this as a natural experiment? Should I collapse them at individual level rather than at the product level?
Thank you,
Veeresh
I have a dataset where there are individuals and there are multiple projects for each individual and I have an external treatment (shock or spam) that has impacted certain projects. So each project has a product owner and multiple team members.
I was wondering how I can use a DID regression in this case. I can create a variable for the product owners who got impacted and pre and post. However, the owners have multiple projects and they may not have been impacted in all the projects. So in the dataset you can see that the external shock occurred in only one product which is apple 12 and hence the owner of the projects is labelled to subject to treatment, however, she/he is not subjected to treatment in the android4 product and hence not sure how to set this as a DID experiment
dataset id product_id owner_treatment post external_shock owner_performance 1 apple12 1 0 1 34 1 android4 1 0 0 54 1 apple12 1 1 1 32 1 android4 1 1 0 21 2 faceb34 0 0 0 345 2 faceb34 0 1 0 454 Any suggestions how I can set up this as DID experiment? And how to do parallel trends? Because there are multiple projects for each individual. So not sure how to set this as a natural experiment? Should I collapse them at individual level rather than at the product level?
Thank you,
Veeresh
