I'm working on a paper by B. H. Baltagi and I am trying to replicate the results. It can be found here, the data is here. I'm supposed to do a linear regression - sounds simple. The author uses Stata, which I have, but I'm more familiar with Gretl.

My problem is, that if I try to perform a regression following the author's procedure on the first page, i.e. choosing all the variables, taking their logs, and so on, I do not get the results that Baltagi gets in Table I on page 3.

Cornwell and Trumbull (1994), hereafter (CT), estimated an economic model of crime using panel data on 90 counties in North Carolina over the period 1981–1987. Table I replicates the Between and fixed effects estimates of CT using Stata.

My idea is (in Gretl) to go to Model -> ordinary least squares and select the variable lcrmrte as dependent variable, and the others, like Baltagi says, as regressors. I cannot see what I do wrong here, but the regression results differ from those in Table I. Some values fit and at least have the right proportion, but others are suddenly (in)significant.

Could this have something to do with the term "between and fixed effect estimates"? What does that mean? I tried to get an explanation online or in the gretl manual, but did not find any. So I got Stata 12, found the xtreg-command, and experimented with xtreg lcrmrte ... fe/be/re, but still did not get the numbers. Does anyone have an idea what I did wrong?

  • $\begingroup$ I'm sorry, I just found the "Cross Validated" Q&A-Site. Should this question be moved there? $\endgroup$ – Marie. P. May 14 '13 at 8:50
  • 1
    $\begingroup$ In stata, you should use xtreg , fe. This is the fixed effects estimator. The between estimate is the same as the fixed effect estimate, but obtained differently. $\endgroup$ – Nameless May 17 '13 at 9:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.