Fixed effect regression

[Svetlana Bondar]

I did fixed effect regression for repeated measures. I included about 10 variables, categorical and continuous, in the list of independent variables. When characterizing my model for F change, p = 0.408 for model 1 and p = 0.416 for model 2. However, when evaluating the coefficients, I found p = 0.038 for one of the independent variables. Is this variable a predictor of the dependent variable, or if F change is not statistically significant, then I cannot further analyze and look for predictors?

[Andrew Musau]

F-change suggests some comparison between nested models. It may be the case that adding a variable or a block of variables does not improve the specification, which is what the F-change is telling you. But for each model, there is an F-statistic which is a test of the joint significance of the variables present. In any case, show us the exact commands that you ran and the resulting output if the preceding is not useful (see FAQ Advice #12 for details).