Introduction to Heteroscedasticity and How to Correct It in Stata

In this video, Dr. Ngozi ADEYELE, PhD, Founder Crunch Econometrix, discusses heteroscedasticity and how to correct it in Stata.

Heteroscedasticity is a statistical problem that occurs when the variance of the residuals (the errors between the actual values and the values predicted by a model) is not constant. This can lead to biased and inefficient estimates of the model’s parameters.

The video starts with an introduction to heteroscedasticity. The speaker then explains how to detect heteroscedasticity using informal and formal methods. The informal methods include plotting the residuals against the fitted values and plotting the residuals against the independent variables. The formal methods include the Breusch-Pagan test, the White test, and the Goldfeld-Quandt test.

Once heteroscedasticity is detected, the speaker explains how to correct it using functional forms, generalized least squares (GLS), or weighted least squares (WLS).

Here are some additional points from the video:

• The speaker recommends using the estat het command in Stata to test for heteroscedasticity.
• The speaker recommends using the log-level, log-log, and level-log functional forms to correct for heteroscedasticity.
• The speaker notes that functional forms are not always the best way to correct for heteroscedasticity. In some cases, it may be better to use GLS or WLS.

I hope this article is helpful!

Credit: To Dr. Ngozi ADEYELE, PhD. Founder Crunch Econometrix