[Stan Breetman]

Stan:
I will focus on the regression models 2 and 4, as they show a higher R-sq within vs regression models 1 and 3.
Eventually, I will select the regression model #4 with default standard errors (all in all, robustified and default standard errors are very similar).

So can you tell me, why i should not use the comand -robust-? Maybe as you stated in another post, so I can give an argument for that . I can find reasons why to do it, but not why to not do it.

[Stan Breetman]

Please do read the material we refer you to carefully. You did not do so with the FAQ (otherwise you would not have posted screenshots) and you did do so not with Allison's blog entry either, which states



If you are asking others to invest their time helping you, you should show that you are willing to invest time as well, not just go for a quick yes or no answer, that you will and cannot not get, anyway. I cannot speak for others but I find this kind of behavior, while understandable, not very respectful.

Best
Daniel

Thank you very much for your important input. In the meantime you were answering me (thank you!), I was looking for for the problem with fixed effects in google. So I try to give my part too.

[Carlo Lazzaro]

Stan:
I was under the impression that, with or without robustified standard errors, your regression outcomes for model 2 and 4 were quite similar: hence, I do not think that heteroskedasticity and/or autocorrelation play a big role in your model.
However, if you think you have a rationale for going -robust-, follow that road.
Obviusly, things might change if you correct your model specification according to Daniel's helpful input about lagged dependent variable.

[Stan Breetman]

With regards to 'better results', ideally, they shall mirror the issues sparked by the data. Moreover, when the model is consistent enough, we get values on the same verge.

The theme relates to core knowledge concerning regression as well as panel data particularities.

Apart from a good book on stats, you may wish to take a look at this thread:

https://www.statalist.org/forums/for...tandard-errors

To end, I kindly suggest to follow the advices from the FAQ.

Among them, the recommendation to share command, output and data (full, abridged or mock). The odds are you will get a much more clarifying answer, provided you follow this suggestion.

Dear Mr. Almeida, dear Carlo

Because of the note one wrote in the thread you suggested, can I proceed like this: 1. I now should just controll for heteroskedasticity
2. If there is heteroskedasticity, I should use -robust- which would be a rational reason. (despite you Carlo think there should not be such issues and thats sure possible, but I think i should test it because I have to argument in my thesis, why I used it or not)
Thank you very much.

The reason for robust standard errors in panel data is because the idiosyncratic errors can have heteroskedasticity or autocorrelation, or both. In Stata's notation, the composite error term is u(i) + e(i,t). Now, pooled OLS leaves u(i) in the error term, which is an obvious source of autocorrelation. But e(i,t) can be autocorrelated. And both u(i) and e(i,t) can both have heteroskedasticity. Clustering handles all of these issues.

In the FE case, u(i) is removed. But e(i,t) is still there. So you still want your inference robust to heteroskedasticity and autocorrelation. If you have been taught that the e(i,t) are supposed to be IID then you have been taught incorrectly.

Hope that helps.

JW

[Marcos Almeida]

The text above, I suppose, comes from a message written by Jeff Wooldridge, a referential author and, fortunately, an active member of the forum.

The text is crystal clear. A great lecture, in a nutshell.

That said, Daniel and Carlo gave insightful replies to the present query as well. Most of the points were approached, if not all of them.

What nobody can replace, I fear say, is the need to take some time to dig deep into the realms of panel data. Be brave and go for it, and the effort may be rewarding.

[Carlo Lazzaro]

Stan:
I wrote that in models 2 and 4, your default standard errors did not seem that different from the robust ones.
Obviously, as Marcos wisely implies, it is up to you to make the final decision and justify it to your audience/reviewers.