Why the p values from my linear regression model don't agree with the ones from ANOVA?

[Yuan Chang]

Hi everyone, here I'm confused: I think ANOVA and Linear Regression Model are supposed to be the same thing, but I got different p values from ANOVA and LRM. The effect size (from ANOVA) and the coefficient (from LRM) are identical, but p values for some variables (e.g. the Black/What gap) are different. I'm wondering what's going on?

[Joseph Coveney]

You might want to read up on what nestreg is doing. In the meantime, try the following and see whether it helps.

version 15.1

clear *
set seed `=strreverse("1474038")'

quietly sysuse auto
summarize rep78, meanonly
quietly replace rep78 = runiformint(`=r(min)', `=r(max)') if mi(rep78)

*
* Begin here
*
regress gear_ratio i.rep78
testparm i.rep78
pwcompare i.rep78, mcompare(scheffe)

anova gear_ratio rep78
pwcompare rep78, mcompare(scheffe)

exit