Random effects model: Run same model, pvalue differed extremely

[Frauke Hildebrand]

Hello,
I run a clustered random effects GLS model.
I have multiple variables. If I run my complete Dofile twice I get two extremely different results for the pvalues of a Dummy variable (0,06 first time, 0.87 second time). I have a dataset of 100 Observations. The Dummy has however only 10 Times a "1".

What is a possible explanation for that? All Other pvalues of other coefficients dont differ.

Thank you so much!

[Carlo Lazzaro]

Frauke:
as per FAQ, posting what you typed and what Stata gave you back (via CODE delimiters, please), will increase your chances of getting helpful replies. Thanks.

[Phil Bromiley]

With 100 observations, on addition to the program, you could give us a copy of the data using dataex. Then, we might have a better chance of diagnosing your problem.