I am analysing panel data and using the plm package in R.
I'm using the plmtest, pFtest, phtest functions which indicate to me that fixed effects model should be used over the pooled OLS and random effect model. However, the R-square for the pooled OLS is 0.73 while the fixed effect is only 0.4. Looking at the prediction, the pooling model is doing better so now I am not sure if I should use the fixed effect model. Any thoughts please?