Twoway scatter with custom line plot from linear regressions

[Dennis Nielsen]

I have performed a linear regression model to investigate the association between a GaitScore and Pain adjusted for age and sex

. regress GaitScore Pain age i.sex

This create the regression model

GaitScore| Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
Pain | -1.030371 .5217701 -3.02 0.004 -2.656475 -0.55933
age | -0.301733 0.201676 -1.53 0.122 -.0005076 0.090161
sex | -2.065706 2.0057716 -0.67 0.480 -.2116815 3.685403
_cons | 49.68492 6.329089 7.85 0.000 37.06195 62.30789
------------------------------------------------------------------------------

If I then plot a twoway scatter with the lfit line
. twoway (scatter GaitScore Pain) (lfit GaitScore Pain)

This draws the scatter with the prediction plots (or how a simple regression linear would look GaitScore, Pain)
However the line I really wanted is the regression line for Pain adjusted for age and sex.

GaitScore = 49.7 + -1.03*Pain

Is that possible, preferable form the stored results from the regression model or manually by writing the formula for the line?

Hope my question is clear..


I am using Stata version 13.1, windows 7,

Thanks in advance

Kind regards
Dennis Nielsen
Biomedical engineer

[Svend Juul]

sysuse auto.dta , clear
twoway (scatter mpg weight)(function y = 39 -0.006*x, range(weight))

[Alan Neustadtl]

The -margins- and -marginsplot- post-estimation commands are very powerful and can do what you want. Unfortunately, I don't know the range and measurement scale) of your data so I can't give you an exact command, but something like the following could work:

regress GaitScore Pain age i.sex
margins sex, at(GaitScore=(0(1)10))
marginsplot, recast(line) noci

Best,
Alan

[Dennis Nielsen]

Thanks so much for all your help.. a special thanks to Svend for some great books on Stata as well.. and thanks to Alan for introducing me to the marginsplot command I have already become quite fond of it... Thanks again..

Kind Regards
Dennis