How can I align samples from two equations?

[sarah_ec]

Hi all!

I want to compare two ols-equations and I am using different variables, however both have the same dated and id-series.

However, not all the explanatory variables are available for all id-series, which means that the number of observations differs within the two equations. So basically eviews creates two different samples, each time using only the observations for which the data is available that is used in the specific equation. To be able to compare both equations, I need to have exactly the same sample. Is it possible to tell eviews that it should only use data for which there values in both equations, without actually using all the variables in each (so that the output table shows the same number of cross-sections and observations)?

Hope my text desribes the problem in an understandable way.

[EViews Gareth]

Use Proc->Make Resid from each equation. That will give you two residual series. Then in your third equation use a sample of "if resid01<>NA and resid02<>na".

[sarah_ec]

Perfect, Thank you so much!