EViews.com

Hello, I am using panel data and am trying to calculate the "between estimator", which as I understand it is the OLS estimator of a regression of individual averages of y on individual averages of x. Is there a canned procedure in the Panel Data estimation box for this? (I have looked in the manual and cannot determine...)

If not, how do I transform my series to be a list of individual averages? Right now my data is essentially in the form below for each series:
indiv 1-1980 = 10
indiv 1-1981 = 12
indiv 1-1982 = 11
indiv 2-1980 = 9
indiv 2-1981 = 10
indiv 2-1982 = 8

If there is no canned procedure I need it to be like this (averages across periods for individuals):
indiv 1 = 11
indiv 2 = 10

Any suggestions for how to generate such a series from existing data or for the canned procedure would be much appreciated. Thanks!!

Hello, I am using panel data and am trying to calculate the "between estimator", which as I understand it is the OLS estimator of a regression of individual averages of y on individual averages of x. Is there a canned procedure in the Panel Data estimation box for this? (I have looked in the manual and cannot determine...)

If not, how do I transform my series to be a list of individual averages? Right now my data is essentially in the form below for each series:
indiv 1-1980 = 10
indiv 1-1981 = 12
indiv 1-1982 = 11
indiv 2-1980 = 9
indiv 2-1981 = 10
indiv 2-1982 = 8

If there is no canned procedure I need it to be like this (averages across periods for individuals):
indiv 1 = 11
indiv 2 = 10

Any suggestions for how to generate such a series from existing data or for the canned procedure would be much appreciated. Thanks!!

You might try creating a new page in the workfile indexed by individual number and then copying series from the panel page, pasting special into the byIndividual page. This can be used to get the mean for each individual in the second page. Then run your regression in the second page.

Thanks much! Because I couldn't figure out how to index by individual I just exported the data to excel and manipulated it there and I'm planning to re-import it and run the regression. However, I have dummy variables in this equation whose average values will be other than 1 or 0 for some individuals. Does anyone know how we treat those in the "between estimator" regression? (This is less a technical support question than a theoretical one, I know...)

Nevermind all, I just included the averages of the dummies and it worked just fine. Thanks for your help Startz!