EViews.com

Hi there!


I am trying to figure out how to conduct an annual cross-sectional logit regression.

I have data for 11 years with a sample size ranging from around 100 to over 1000.

As far as I know, the usual way of conducting a cross-sectional analysis is by creating a (balanced) panel workfile.
Unfortunately, the data they I have in the form of Excel spreadsheets is not suited for conducting the tests in this way.

Therefore, I thought about creating one undated workfile for every sample year with no. of observations = no of sample constituents and then estimating the equation using logit as the binary estimation method.

The excel spreadsheet that I would import would look like this:
dependent variable explaining variable 1 explanining variable 2 explaining variable 3
company 1 0 0.3 0.2 0.1
company 2 1 0.8 0.65 0.56
.
.
.


Having done this for the whole sample period (11x), I could then calculate the average estimations for each of the variables I am looking at.


Is this a feasible way to conduct such an analysis!



Thank you!


Regards,

MS

If you're not doing panel analysis (and your post seems to indicate you're not), then there is no need to create a panel workfile. Creating a page for each year is perfectly fine, so I think you're doing the right thing.