0

I am looking for a reference to perform variable selection with panel data. I have 9 variables and want to test which variables to keep or remove. Someone recommended step-wise analysis, but I cannot find a resource for step-wise analysis with the panel data. I am currently working in R.

AData <- pdata.frame(Data1, index=c("Recipient","Year"))
random <- plm( y ~ x1 +  x2  + 
             x3 + x4 + x5 +
             x6 + x7 + x8 + x9,
           data=AData, model= "random")
summary(random)
kjetil b halvorsen
  • 63,378
  • 26
  • 142
  • 467
Bonnie
  • 1
  • 1
  • 1
    On this site stepwise is not popular ... see https://stats.stackexchange.com/questions/258026/are-there-any-circumstances-where-stepwise-regression-should-be-used – kjetil b halvorsen May 24 '21 at 01:03
  • Can you clarify what the aim is here? There's no test for whether a variable should be kept or removed: only theory can tell you whether predictors make sense. If you just want to know which ones are most "important," then compare the standardized coefficients. If you absolutely need to cut down predictors, then you could compare information criteria between models with all versus some predictors – Billy May 24 '21 at 01:40

0 Answers0