Hi all,

I performed ROC regressio using the rocreg command and I have the desired results. I know that the significant variables affect the cases, controls and ROC curve respectively. However, I'm not sure how to interpret the B. coefficients. What exactly does the B. coeficet of 1.4 mean in the ROC model in the attached figure?

Hier is what I ran
rocreg prevcase corscore [pw=pweighting], ctrlcov(sc_bmi sexcoded basengtswt sc_flu basehemopt smkhst) roccov(basecough) ctrlmodel(linear) cluster(case_number) probit ml nolog


* Example generated by -dataex-. For more info, type help dataex
clear
input float(prevcase corscore sc_bmi) byte sexcoded float basengtswt byte sc_flu float basehemopt byte smkhst float(basecough pweighting)
1 92.64069 25.47 0 0 1 0 0 1 1.263389
1 89.17749 17.23 1 0 0 0 1 0 1.263389
1 27.70563 17.77 1 0 0 0 1 0 7.919843
1 76.62337 23.81 1 0 0 0 1 0 1.263389
1 73.59307 22.57 0 0 0 0 1 0 1.263389
1 64.50217 20.18 1 0 0 0 1 0 1.263389
1 90.47619 17.79 1 0 0 0 1 0 1.263389
1 87.44588 17.8 1 0 0 0 1 0 1.263389
1 9.95671 17.38 1 0 0 0 1 0 7.919843
1 92.10526 23.29 1 0 0 0 1 0 1.263389
end
label values smkhst YesNo
label values sc_flu YesNo
label values prevcase YesNo
label values basecough YesNo
label values basehemopt YesNo
label values basengtswt YesNo
label def YesNo 0 "No", modify
label def YesNo 1 "Yes", modify
label values sexcoded gendercode
label def gendercode 0 "Female", modify
label def gendercode 1 "Male", modify
label var prevcase "PrevalentTB"
label var corscore "RISK11 Score"
label var sc_bmi "BMI"
label var sexcoded "Gender(M)"
label var basengtswt "NightSweats"
label var sc_flu "Flu-like"
label var basehemopt "Hemoptisis"
label var smkhst "History of smoking"
label var basecough "Cough" Array