Mixed effect - Pooled ols Different results interpretation I have a question. I have collected data regarding the performance of companies and their board structure. I want to find the effect of the Board structure upon the performance and I am using pooled OLS and mixed effect for that. I know that the two methods differ but the results are also different. in my OLS results some of the variables significant and a few not. When i run my mixed effects some variables lose their significance and other gain stronger. The sample is the same. Any idea of how can i explain this thing?
 A: I can think of at least two related explanations:


*

*If the assumptions of the pooled OLS were appropriate, you should expect to get the same conclusions from the pooled models as from the mixed model (assuming that both are correctly set up). Consequently, that you get such different answers can be thought of as meaning that the pooled OLS model should not be used. (IMO pooled OLS is one of those models that is only taught for pedagogical reasons, and should not really be used.)

*You have not completely understood the within board versus between board effects in your mixed model. That is, the extent to which a board's composition over time influences performance is the within-board effect, whereas the effect to which board compositions between companies effects performance is the between board effect. Once you understand these effects, you should be able to reconcile them with the pooled OLS, as the pooled OLS will basically be showing you an average of the within and between effects, provided there is not too much else going on in your model.
