Unit Root Test / Lags!

[schaich]

Hi,

I am trying to run Augmented Dickey Fuller test to my series (286 obs, monthly data) but I'm not sure how to manage the lags. The default criteria keeps changing when I use trend, intercept or none, and I can swear I got different lags last time I did the same test. I'm sure I'm doing something wrong. Sometimes I get a I(1) series and others a I(0) (must be I(1)). Results also change when I test PP and KPSS.

Is there a criteria to choose a correct number of lags? When is PP and KPSS relevant?

If somebody would be willing to help by corunning the test I would very grateful, just don't know what else to do.

Thank you!
Maria