Extract Std.Err. and 95% Conf. Interval after regression

[Junran Cao]

Hi Statalist

After running a regression, I understand we can extract the estimated coefficients in the following manner

. use http://www.stata-press.com/data/r16/mheart0.dta
. regress bmi age smokes female

Once the model is estimated, we can type

ereturn list

to see the list of results that can be extracted individually. In this case,

matrix list e(b)

would give us the estimated coefficients.

My question is can the same thing be done for the estimated standard errors and confidence intervals? The matrices e(se) or e(ci) do not appear to be there.

Thanks.

[Wouter Wakker]

mat list r(table)

will give you the full table of coefficients with se, ci, p-value etc.

[Maarten Buis]

Alternatively see this Stata tip: https://www.stata-journal.com/articl...article=st0137

[Junran Cao]

Thank you Maarten for that excellent article!
And thank you Wouter for letting me know about -r(table)-

Regarding -r(table)-, can I please ask a follow up question on how to index a row of values from that matrix? Specifically, to retrieve a row vector of SE like that of e(b).

I have tried

. matrix define newmat = r(table)[2,]

and similar commands but they weren't correct.

Thanks

[Wouter Wakker]

You could do that in this way.

mat define temp = r(table)
mat define newmat = temp[2,1...]

where the first two dots mean 'till' and the last dot means 'last'. See also

help matrix extraction

[Junran Cao]

Thank you very much Wouter! Much appreciated.

For anyone else who may have had the same question as I did and found this tread, after tying in

help matrix extraction

as per Wouter's post above, page 350 of the documentation (Stata 16) contains detailed information on this topic.