Post estimation tests of Panel data using Stata: VIF: Serial Correlation: heteroskedasticity

In this video we conduct post estimation tests on a panel data in stata such as VIF for multicollinearity, Breusch-Pagan / Cook-Weisberg test for heteroskedasticity, Ramsey RESET test for omitted variables, Breusch-Godfrey LM test for autocorrelation.
I have conducted three tests for serial correlation errors, like breush-godfrey test, durbin-watson test and through uhat test. You can choose any one out of these as well.
#statastics : #panel #datascience : #correlation #analysis #fixed #effects AND #random #effects : #HAUSMAN #test #hausmannskost #stata #paneldata #correlation #fixedeffect #randomeffects #random #dataanalysis #paneldata #vif #serialcorrelationerrors #heteroskedasticity #postestimation #dataanalysis #paneldata #random #correlation #fixed

[@liz_f1]

Thank you very much! This video is very useful, I've been looking for this information for a long time....

[@GiulianaMacari]

Hello. Am I correct in saying you run these diagnostic tests on the simple OLS regression that has not fixed time and entity effects? Then, if the data passes all of the tests, you can go on to conduct the fixed effects regression?

[@am.am16]

Grateful!

[@damir2go]

Can we do this Postestimation tests for xtreg commands because i need to run the tests on my fixed effects regression

[@dr.rahman107]

Hi, Thanks for the question. all tests not possible with xtreg. If you go to the post estimation tab, you can find that some of the tests only possible with time series setting. So, for that you have to change your data to time series if you need those tests, such as Breuch-Godfrey test. Others you can do with xtreg or the panel data setting.

[@damir2go]

@@dr.rahman107 thank you for answering fast and precise. how do i find out if i need postestimation Tests for my paneldata or Not? Its Hard to find a source with a holistic explanation for my analysis

[@dr.rahman107]

@@damir2go . You should do post estimation tests to prove that your data and results does not suffer from any biases. However, which all tests is subjective. Some do both tests related to time series and panel even if your hypothesis testing is only based on panel set up. Better you can read research papers in your field and in journal where you are aiming to publish. that can give you a hint about what all post estimation tests you should conduct.

[@farhanfarzam4278]

Thank you sir for your insightful lessons, i want check for heteroskedasticity but i go this error "estat hettest not valid
r(321); " pls pus some light on it.

[@anjalitandon4047]

Thank you for the informative video. Could you please share how correct if the problems of heteroscedasticity, autocorrelation, multicollinearity are found in panel data

[@dr.rahman107]

thanks for the question. If I understood your question correctly. if the significance level of these tests is greater than 0.05, then there is no problems. If you find these tests significant(smaller than 0.05), your data has these issues, and hence, your results will be biased. Hope it answers the question.

[@EasyLearningMMA]

@anjalitandon4047 maam can you share any link from where you get solution of these problems? as If you these tests significant(smaller than 0.05), is only indicator .. not the solution....

[@AndiHamzah-zp6cv]

This is insane, you didn’t do panel violation diagnostic tests, what you just did is only time series violation assumptions, whatch out your title please, it’s totally different panel and time series

[@dr.rahman107]

Hey, go and do whatever you think is right. Or make a video and post of what you think is right. Who stops you. I did what I understood during my research and tried to help people. Cheers.

[@prajwalkoirala1821]

@@dr.rahman107you are misleading people here, you just conducted a Pooled OLS disregarding any cross sectional variation (which is the main motive of panel data study). It would be better if you changed your title to Pooled OLS diagnostic test in panel dataset. Nonetheless, thankyou for presenting a detailed diagnostic analysis of Pooled OLS

[@elvirc5317]

@@dr.rahman107all those tests are informative as data have such a problems in the beginning and we have to deal with problematic data and post analysis is only possible solution (after ols), a lot of people do not understand the logic of initial testing in order to choose right estimator. thanks for video