Hello everyone,
I am running a panel data regression model in which I observe 4 time periods and 66 observations, using a fixed effects regression and I am currently including the I.year variable to control for year fixed effects. However when I run the regression without i.year, results vary significantly and coefficients that were statistically insignificant now become significant. I ran a testparm i.Year test in order to see if there was need to include the i.Year variable and the results demonstrated that yes. I was just wondering if there was any reasons why I would not include them or if it Is always wise to include i.year?
Many thanks
Pepito