Skip to content

random effect

2 messages · Joana Martelo, Joerg Luedicke

#
Joana Martelo asked:
"However, do you think that considering year a fixed effect will
affect the relationship between the other explanatory variables and
the response?"

Entering an additional variable in a regression model will always
change the effects of other predictor variables on the outcome that
were estimated in a previous model. In your case, entering year as a
binary predictor essentially means that you constrain your other
effects to be the same in both years. If you have reason to believe
that the effects are different across the two years, you can interact
year with your other predictor variables and thereby relax the
assumption that the effects are the same in both years.

Joerg
On Sun, Oct 7, 2012 at 11:29 AM, joana martelo <jmmartelo at fc.ul.pt> wrote: