RE: Twoway scatter command for logistic S curve

[DY Kim]

Hi! I am trying to create a logisitc curve based on the estimated logit model. The twoway scatter command generated the attached graph. twoway scatter yhat rearrest age, connect(1 .) symbol(i O) sort ylabel(0 1)

However, the probability dots/lines did not look like a logistic S curve. Is there anyway I can create a logistic S curve based on the estimated model.

Thank you for your help!

[Andrew Musau]

Graph the predictions:

webuse lbw, clear
logit low age lwt smoke ptl ht i.race
predict xb, xb
predict pr, pr
tw line pr xb, sort xtitle("{&beta}'x") ytitle("probability") scheme(s1color)