[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
Re: st: Regression p-values
Sounds like you don't have many degrees of freedom left. Looking at Mood, Graybill & Boes (1974), I see that a t(df=5, p=0.10) = 2.015 -- so you must have something close to that in terms of df. The robust & cluster options will further use up degrees of freedom which I guess is one of the costs of using it.
Hope this helps. You might want to paste in your output (the relevant parts) in a subsequent posting if it appears I'm not getting your inquiry.
>Why is it that sometimes the regression outputted p-values don't match
>up with the reported t-stats? For example, t=1.96 should correspond
>to p = 0.05. But in current output using the robust and cluster
>options I have one stat with t=2.02 and p=0.078.
Eric G. Wruck
2535 Sherwood Road
Columbus, OH 43209
* For searches and help try: