How to extract regression model p-value (Prob > F) and store it in a macro?

[Dusan Petrovic]

Hello statalisters,

This question has already been probably asked, but I couldn't find the way to figure it out, sorry.

My problem is that this p-value is not stored in the postestimates (e(...)) nor the r(table). I've also found that there is this parmby parmest package that I installed, but

parmest, list(p)

is not helping me (I could have this with r(table)).

Thank you for your kind help

[Rich Goldstein]

the p-value is stored in r(table) - it is in row 4 and marked as "pvalue"

[Dusan Petrovic]

Thank you, but I want actually is the (prob > F) => the probability which is associated with the F statistic displayed in the upper right corner below F(df1, df2) and above R-squared.

[Rich Goldstein]

yes, sorry about that - I tried to cancel it before it posted but, of course, failed

note that the numerator df, the denominator df and the value of F are all in -ereturn list-; you could then use the appropriate prob density function (see -h function-) to get the p-value

[Dusan Petrovic]

Great, thanks Rich. Btw it's

 Ftail(df1,df2,F-stat)

for the ones who want to know.

[Maarten Buis]

Also see: M.L. Buis (2014) "Stata tip 116: Where did my p-values go? (part 3)", The Stata Journal, 14(1) pp. 218-220.

[Dusan Petrovic]

Thank you Maarten, I'll check it out

[Luke Sonnet]

Sorry for bumping an old thread, but for those looking, it's actually part 2 of the "where did my p-values go" that has the direct answer to this question (F-test pvals after regression)

Here's the link: M.L. Buis (2007) Stata tip 112: Where did my p-values go? (part 2)

And the code:

Ftail(e(df_m), e(df_r), e(F))