classifying data as a panel however there are year duplicates that are relevant

[Amina Abubakar]

Hello,

I am trying to classify a data set into a panel, however the issue is that I have repeated time variables. This is because I am working with a dataset that gives information on different projects within the same year (eg 2008 appears multiple times for each country because multiple projects occur). Is there any way I can circumvent this and turn my data into panel without having to drop off those observations?

Any assistance would be very much appreciated.

[Clyde Schechter]

The short answer is no.

But probably you don't need to, anyway. Unless you are going to do analyses that rely on lags, leads, seasonal differences, etc., or estimate autocorrelations, you can just run -xtset country- with no time variable and then everything else will be fine.

If you were planning to do analyses that rely on lags, leads, autocorrelations, etc., then you need a new plan. Because it's not just that Stata won't do those analyses in this kind of data, it's that such analyses are not even mathematically possible. Think about it. If the same country has more than one project in 2008, then which one of those would be "the lag" of that country's observation in 2009? It's impossible to say. Such constructs are undefinable in the absence of true panel data structure.

[Jared Greathouse]

You need a variable for each project, by the sounds of it

[Amina Abubakar]

Thank you Clyde, I am not planning to use lags or autocorrelations and I have been able to go on just fine with your suggestion of running 'xtset country' thank you very much for you assistance