Heteroskedasticity ?

[fahmi mutassem]

Does this show rvf plot above show Heteroskedasticity ? I ran :

estat imtest, white

White's test for Ho: homoskedasticity
against Ha: unrestricted heteroskedasticity

chi2(19) = 22.76
Prob > chi2 = 0.2479

which shows no homoskedasticity , but I am still not sure.

[Marcos Almeida]

Welcome tp the Stata Forum/ Statalist.

Please, do read the FAQ. There you’ll find advice about the best way to post a question, including how to share data/command/output/graphics.

[Carlo Lazzaro]

Fahmi:
the White test tells you that, given your data, there's no evidence of heteroskedasticity (the usual caveat sounds: absence of evidence is not evidence of absence; see https://www.ncbi.nlm.nih.gov/pubmed/7647644).
As far as your plot is concerned (by the way: it is always a wise habit to visually inspect residuals distribution, over and above the outcome of analytical tests), you may have some nasty behaving residuals near 0, but those are probably not strong enough to reject the null.
As an aside, Marcos' helpful recommendation obviously applies.

[Marcos Almeida]

I agree with Carlo. By the way, with the exception of a residual beyond -2 in the y axis, the plot is somewhat nice. That said (and I believe this is not the case), when the statistical test and plot diverge in terms of results, I tend to prefer the second.