Fixed Effects, Random Effects, First Differences

[Charlie Austin]

Hello everybody. I am an undergraduate economics student struggling with two main concepts related to fixed effects, random effects and first differences tests. I would be beyond grateful for anybody's insight into my questions. I imagine they are relatively easy to answer.

1. Are FE, RE and FD all used to diagnose/ get rid of omitted variable bias? If not, what else are they used for? If so, what is the difference in use between the 3? I know that for FE and FD, they are the same exact tests if your T value = 2, but if T>2, it is better to use FE. What kinds of problems would be diagnosed by FD and not be FE or RE, and vice versa?

2. How do I run these tests on STATA? I have looked online and have struggled to find detailed answers on how to run these tests. And, once I run these tests, how do I interpret the results to improve the accuracy of my data?

Again, I would really appreciate explanations for these problems. Thank you!

[Carlo Lazzaro]

Charlie:
welcome to this forum.
I find your post a bit confused since what you mention in 1. are actually estimators, not tests.
I would recommend you to grasp the backbones of panel data econometrics via one of the many valuabe texbooks that you can find at the library of your university.
Stata users dealing with econometrics find https://www.stata.com/bookstore/micr...metrics-stata/ a valuable mix of theoretical and empirical stuff.

[Charlie Austin]

Thank you, kind sir.

[David Benson]

I tend also to use Jeff Wooldridge's book "Introductory Econometrics" https://www.amazon.com/gp/product/0324113641 for that (your questions are answered in Chap 13 & 14. Although his book doesn't provide the relevant Stata code (whereas the Microeconometrics Using Stata book that Carlo noted, does).