Date:         Sun, 24 Jul 2011 17:49:22 -0400
Reply-To:     Anna Tsui <ascui@UIC.EDU>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Anna Tsui <ascui@UIC.EDU>
Subject:      Hausman test in proc panel

Hi there,

I am running proc panel with 100 people and 4 years of data. I asked for
Hausman test to compare random versus fixed effects, but sometimes I get
the following warning:

"WARNING: Hausman statistic cannot be calculated because different
variables were dropped in the random effects model than in the fixed
effects model."

I think it happens when I include a variable that does not change across
years, i.e., it only varies across people. Do you think this is the
reason? If so, what can I do to get the Hausman test?

Thanks a lot in advance.

Anna