Home > Statistics Every Writer Should Know > The Stats Board > Discusssion
is ROC correct?
Message posted by Jen (via 155.42.42.91) on July 3, 2001 at 1:36 PM (ET)
Hi, I am trying to determine the optimal length of time for a correctional program. I have a cross sectional population with 100 participants who have been in the program for various lengths of time. I am using 2 continously measured scales. So can I use a ROC to determine after "N" months scores no longer increase significantly? If not, what do I use?? Please feel free to email me at tragerjs@netscape.net...
READERS RESPOND:
(In chronological order. Most recent at the bottom.)
Re: is ROC correct?
Message posted by JG (via 128.8.23.44) on July 3, 2001 at 4:59 PM (ET)
What is ROC ? What type of correctional program are we talking about ? What scales are you using ?
Re: is ROC correct?
Message posted by Jen (via 155.42.42.91) on July 4, 2001 at 2:04 PM (ET)
I am refering to receiver operating characteristic curves (ROC curves). It is a cognitive restructuring program that will be evaluated using the Criminal Sentiments Scale - Modified and another brief assessment scale that involves interviewing the group facilitator, with a possible range of scores from 0-40. Would this have something to do with nonlinear regression? But I want more than just an equation, I would like the point on the regression line/curve that decreases in score are no longer significant. In other words, perhaps there is no significant decrease in scores after 4 months in the program because scores either level off or start to rise again.
Re: is ROC correct?
Message posted by JG (via 128.8.22.198) on July 4, 2001 at 6:47 PM (ET)
The operating characteristic curve simply plots the alpha value for a test as the relavent parameter changes.
However, receiver OC and cognitive restructuring are terms that I am not fomiliar with.
Your $5 contribution helps cover part the $500 annual cost of keeping this site online.