Suppose I have a multiple linear regression model and the errors are not distributed normally.Does the central limit theorem hold true in this case? should i trust the p values of the coefficients of variables assuming the residuals are not skewed?
Will really appreciate any help on this, thank you :)
