Great video Sebastian! Explained in detail and easy to understand. Might i also add the estat hettest command, as a way to reject or not, the constant variance null hypothesis.
Many thanks for this video. I'm just wondering, can we just use the command ( hettest) after the original regression ( without typing "robust" at the end of it) to figure out the presence of heteroskedasticity?
Hi, thank you for this helpful video! When I followed the first two test, the p-value for my f statistic was significant for heteroskedasticity (P= 0.0591; P=0.0024). However, when I ran the 3rd test in the video, my p values were no longer significant (p=0.2514). Not sure what this means. Should I follow you're correction to the heteroskedasticity video anyway or is there something more fundamentally wrong with my model?
I am surprised your fully specified White test would have such a high p-value compared to the others. It's possible you made a mistake with the test, but I'm not sure. There could be other problems with your model, but these tests don't provide any information on that. In practice, most economists just use White's robust standard errors on all regressions.
I have a question when I’m doing my white test I got a result of 193 degrees of freedom that’s exactly the same as the dataset. Is there a way to fix it?
You need to use heteroscedasticity robust standard errors (I have a video on this) and should consider first-differences (shown in my fixed effects video).
This gives us a nonlinear interpretation, such that we're looking at proportional changes, rather than level changes. It's doesn't really have anything to do with heteroscedasticity.
You need to add terms for all combinations of variables. For example, if you had three variables: a, b, and c, then you need: a^2, b^2, c^2, ab, ac, and bc.
@@sebastianwaiecon Thank you so much. I got it. Could you please send me your gmail or email address which you use frequently? I have got many questions to ask you about Stata. I really need help. Thanks in advance
After regressing e2 on yhat and yhat2 , I obtained this F(2, 1327) = 0.43 Prob>F = 0.6534 Does this mean that I don't have a heteroskedasticity problem in my model?
when i use BP test of heteroskedasticity and i find that p-value =0.000 so there is not a heteroskedasticity so we don't need to do the white test?? that's it?
@@sebastianwaiecon I think I'm just gonna stick to Breusch-Pagan and check a plot of my residuals! I tried a few custom packages, but they were wack. Thanks tho
Sir, I need a suggestion. My 1st White's test -----------------------> Prob > F = 0.0704 , F(7, 648) = 1.88) 2nd White's test ------------------------------> Prob > F = 0.1626 , F(2, 653) = 1.82 Should I accept that there is "Heteroscedasticity" ??? or not? Please guide me.