Low within variation independent variable and fixed effects

[Mimina John]

Dear all,

I am running a regression where one of my (independent) variables of interest has low within group variation. I have reasons to believe however that I should impose fixed effects. After doing some reading I read that estimates can be imprecise as standard errors can become too large. Does this have a direct impact on the p-values however. And if so, do p values become too large to too small?

Thank you


Edit: my intuition would say that these fixed effects are likely correlated with these variables and therefore p values would become higher.

Edit 2: I meant these fixed effects might pick these effects up

[Clyde Schechter]

If your variable has very little within-group variation, then the standard error for the coefficient of that variable will be large. That has a cascade of effects: the z or t statistic has that standard error in the denominator, so large standard error entails a small test statistic. Small test statistic implies large p-value.

[Mimina John]

Thank you very much. That conforms my suspicions