heteroskedastic multiple regression

[Rev Payaaru]

Hello everyone,

Considering, using Hetroskedastic multiple regression, as the residual vs fitted graphs is in funnel shape for the multiple regression. Only one variable is found to be violating homoscedasticity. While the coefficients and p-values from both the results appear, wondering whether the choice of Hetroskedastic multiple regression is right.
Looking for some papers to understand how the results are interpreted and presented. Could not find any papers on Google Scholar. Appreciate any pointers.

Thanks!

[Richard Williams]

Are you thinking of using the hetregress command? The Stata manual entry for hetregress has some references. Along with examples and explanations of its own.

[Rev Payaaru]

Thank you for the reply!
Read the Stata 15 manual. For example, if I have to understand how to interpret these results, I am looking at some cited work to learn.

. hetregress food_exp income, het(logincome)

Fitting full model:

Iteration 0: log likelihood = -227.889

Iteration 1: log likelihood = -226.61039

Iteration 2: log likelihood = -225.72188

Iteration 3: log likelihood = -225.71519

Iteration 4: log likelihood = -225.71519

Heteroskedastic linear regression Number of obs = 40

ML estimation

Wald chi2(1) = 135.11

Log likelihood = -225.7152 Prob > chi2 = 0.0000