Sorry, just read about not posting the same question in more than one forum. Please ignore this one then.

======================

I have a sample of 150 from a population of 300, so when deciding whether a variable is significant in a regression I need to use a finite population correction factor. I know how to do this when calculating confidence intervals or t-tests, but I'm not sure how to in regression.

The finite population correction factor I have is sqrt((N-n)/(N-1)).

I am using SPSS and I think I need to change the stepping method criteria in the options dialog in linear regression. I can either change the probability of acceptance (currently entry = 0.05, removal = 0.1) or the F value.

Any ideas?

Thanks.

======================

I have a sample of 150 from a population of 300, so when deciding whether a variable is significant in a regression I need to use a finite population correction factor. I know how to do this when calculating confidence intervals or t-tests, but I'm not sure how to in regression.

The finite population correction factor I have is sqrt((N-n)/(N-1)).

I am using SPSS and I think I need to change the stepping method criteria in the options dialog in linear regression. I can either change the probability of acceptance (currently entry = 0.05, removal = 0.1) or the F value.

Any ideas?

Thanks.

Last edited: