I am running a fixed-effects model of the type "between" using the plm package in R. According to the Breusch Pagan test heteroskedasticity is present. However, th