[Clyde Schechter]

I will also be doing a fixed effects regression which as I understand will eliminate the attrition problem since it is measuring within individuals.

That's not correct. You can still have an attrition bias: the within-individual relationships among the variables may be different for completers than they are for dropouts.

rlink m:1 panelid, link(participation_years)

gave me the error: option frame() required. I changed it to the following:

frlink m:1 panelid, frame(participation_years)

which seemed to work and did the job you described.

Sorry, that was my mistake. Glad you found the correction.

[Komal Jain]

Hi,

I have a repeated cross-section data of 1,81,000 observations.

My data is in long format where each respondent (farmer) has multiple rows if he participated in multiple seasons. For example: if he participated in Rabi 2017 and Rabi 2018, he would have 2 rows.

There are some farmers who were interviewed in every season, some interviewed only in 1 or more seasons etc.

Now my aim is to find: how many new farmers were interviewed every season (there are 10 seasons).

Can you please guide me how to do this?

Thanks in advance,

[Nick Cox]

Something like

egen when_started = min(season), by(farmer)
egen tag = tag(farmer)
tab when_started if tag

would show you a tabulation of farmers by their first season in the database. Naturally your variable names may well be different.

[Clyde Schechter]

It is possible that some farmers started their participation in the first season, but then did not continue through the tenth, or perhaps participated only intermittently. So an approach that would exclude these farmers would be:

isid farmer season, sort
by farmer (season): gen byte completer = (_N == 10) if _n == 1
count if completer

[Komal Jain]

Thanks a lot