Date: Tue, 12 May 1998 09:05:54 -0600
Reply-To: "A. PAUL BEAULNE" <A.BEAULNE@V-WAVE.COM>
Sender: "SPSSX(r) Discussion" <SPSSX-L@UGA.CC.UGA.EDU>
From: "A. PAUL BEAULNE" <A.BEAULNE@V-WAVE.COM>
Subject: Re: ROC curves for yes/no response
In-Reply-To: <s5581ded.032@harthosp.org>
Content-Type: text/plain; charset="iso-8859-1"
I am going to take a stab at this but I am not an expert in this form of
analysis
My understanding of ROC's is that one uses
different cut off points along a scale to make decisions
For example the SAT scores that are employed for admission to university
or a scale to decided if a particular treatment or diagnostic should be
done.
With each cut of score there is an associated sensitivity (true positive)
and specificity (true negative).
By employing several cut off scores at which a decision is made one then
will
have a number of pairs of S&S measures.
Plot these pairs will give the ROC curves
Decisions with respect to the "Best" cut score can be made through analysis
of the ROC curve
Therefore I have no way of understanding how your question can be answered
given what I have just said
> -----Original Message-----
> From: SPSSX(r) Discussion [mailto:SPSSX-L@UGA.CC.UGA.EDU]On Behalf Of
> Jeff Mather
> Sent: May 12, 1998 8:01 AM
> To: SPSSX-L@UGA.CC.UGA.EDU
> Subject: ROC curves for yes/no response
>
>
> 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
>
