Date:         Wed, 13 May 1998 13:01:37 -0400
Reply-To:     sbaker@nt.dma.state.ma.us
Sender:       "SPSSX(r) Discussion" <SPSSX-L@UGA.CC.UGA.EDU>
From:         Stephen Baker <stephen.baker@UMMED.EDU>
Organization: Massachusetts Division of Medical Assistance
Subject:      Re: ROC curves for yes/no response
Comments: To: Jeff Mather <Jmather@HARTHOSP.ORG>
Content-Type: text/plain; charset=us-ascii

The ROC analysis is generally used to evaluate the inherent predictive value of
an instrument for a dicotomous outcome.  If the instrument is dichotomous the
analysis becomes meaningless, the appropriate analysis is a
sensitivity/specificity analysis, i.e # true positives/(true positives + false
negatives) and # true negatives/(true negatives+false positives).  Since your
instrument I understand is this sort.
 
Jeff Mather wrote:
 
> This is a bit off the SPSS path, but can anyone tell me if there is any way
> to use a nominal response to draw an ROC curve?  I'd like to compare
> ultra-brief (one question) instruments for a diagnosis of depression (yes
> or no) with standard questionnaire that uses a cutoff on an interval scale.
>
> My assumption was that this could not be done with a y/n response; that
> you need at least a three point ordinal scale (ie. always, sometimes,
> never) to be able to draw the curve.
> thanks
 
 
 
--
Stephen P. Baker, MScPH                          (617) 210-5695
Director, Research and Evaluation                (617) 451-7105 fax
Massachusetts Division of Medical Assistance     sbaker@nt.dma.state.ma.us
600 Washington Street
http://www.ummed.edu/pub/s/sbaker
Boston, Massachusetts 02111