scatterplot of regression including interaction

[Cameron Brick]

Hi all, thanks for reading. I understand how to graph a scatter plot between two variables using 'twoway'. However, how do I generate a scatter plot including an interaction term and covariates? I searched here and didn't see a solution yet. I'm new to the forum, so I may have missed it: apologies if so.

regress flossesused age gender i.ethnicity pastflossweek c.cond2##c.ce2
margins, at(ce2=(.45 1) cond2=(0 1))
marginsplot

That's the code I'm using now, simplified for this post. The graph doesn't yield scatter, even if I call 'marginsplot recast(scatter)': in that case, it just gives the point estimates of the margins and doesn't show scatter or a fitted line.

[Kim Mannemar Sønderskov]

I'm not sure what you mean by "a scatter plot including an interaction term and covariates", but perhaps the option -addplot(scatter y x)- is what you are looking for.

Assuming you want to show the predicted relationship between ce2 and flossesused over the range of ce2 (.45 to 1) for cond2 = 0 and 1 and include the scatterplot of ce2 and flossesused, you could go like this:

regress flossesused age gender i.ethnicity pastflossweek c.cond2##c.ce2
margins, at(ce2=(.45 1) cond2=(0 1))
marginsplot, noci recast(line) addplot(flossesused ce2)


You could also consider using contour plots.

Best,
Kim

[Cameron Brick]

Dear Kim, thanks for your suggestions. I found success with this code:

marginsplot, noci recast(line) addplot(scatter flossesused ce2)

In your code section 'scatter' was dropped. This is resolved. Thanks!