Graphing confidence intervals for twoway scatter

[Jonas Wereld]

Following this tutorial, one can graph:

logit hiqual avg_ed 
 predict yhat1   
 twoway scatter yhat1 hiqual avg_ed, connect(l i) msymbol(i O) sort ylabel(0 1)

I would like to graph the confidence intervals as well. I know of the lfitci option for graph twoway, but couldn't find anything for twoway scatter. How would one proceed here?

[Nick Cox]

You have to use predict again to get appropriate variables for upper and lower bounds. There is no way that scatter understands where what it is plotting came from.

[Jonas Wereld]

Thanks. Following your suggestions, I followed this website: http://www.stata.com/support/faqs/st...nce-intervals/
I now have two variables, plb and pub which are the lower and upper bounds. I am wondering however, how can I now integrate these variables in the twoway scatter graph?

[Nick Cox]

They are just more variables to show with lines. If you want a smoother curve, you may have to jump to twoway mspline

[Mathias Pedersen Heinze]

I am wondering however, how can I now integrate these variables in the twoway scatter graph?

On a very basic level, you can just add a second plot to the twoway, say rcap for instance.

twoway (scatter yhat1 hiqual avg_ed) (rcap plb pub avg_ed)