which regression model to use

[Wumbe Emma]

I am trying to answer an objective which looks at: factors associated with perinatal mortality before and after introduction of medical drone delivery services. What type of regression model will be more appropriate? Would i need to take out those with the outcome before the intervention and run the regression model and then do same for those with outcome after intervention?

[Luca Calianno]

What kind of dataset do you have (i.e. unit of observation, time period, it is a panel...)?

Do you have the control group (units that did not receive the intervention)? In this case you could run a Dif-in-dif:

 reg perinatal_mortality i.after##i.treated

where after = 1 if after the intervention and treated = 1 if treated.

Without control group I would just run something like this but it is more difficult to argue that the difference in the post period that you observe is due to the intervention.

 reg perinatal_mortality i.after

[Carlo Lazzaro]

Emma:
as an aside to Luca's hepful advice, if you want to go DID, you may want to consider -help didregress-.

[Wumbe Emma]

I do not have the control group. I intend looking at the difference in the two time points

[Carlo Lazzaro]

Emma:
Luca's second code sounds good, then.

[Wumbe Emma]

Thanks for the support