Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Muhammad Riaz <riaz_stata@live.co.uk> |
To | <statalist@hsphsun2.harvard.edu> |
Subject | st: Difference between the results of lroc after logistic regression and roccurve/rocreg with covariate adjustment |
Date | Sun, 16 May 2010 11:02:18 +0100 |
Dear all, I am adjusting for the covariates effect in ROC curve analyses. I was thinking to use logistic regression with covariates entered into the model and then use the lroc command to get the adjusted estimate of AUC and ROC curve. However the google search introduce me to roccurve and rocreg commands and correct methology for covariate adjustement in ROC analyses. I used both of the commands (i.e lroc and roccurve) and got very different values for area under the curve. (i.e. 0.85 from lroc after logistic regression with covariates in the model, and 0.52 from rocreg (coeff: of alpha_0=0.52)) My understanding is that "lroc" is measuring the predictiveability of the logistic regression and rocreg measure the predictiveability of the marker in the presence of covariate effect. However I was thinking that the lroc result of logistic regression without any covariate in the model should be equal to the roccurve/rocreg but it is not? AUC values are 0.68 and 0.61. Here i dont understand what could be the reason? could there be something wrong with roccurve/rocreg command which i dont understand I will really appreciate your comments and feedback Thank you in advance, Muhammad _________________________________________________________________ http://clk.atdmt.com/UKM/go/197222280/direct/01/ Do you have a story that started on Hotmail? Tell us now * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/