Discriminant Function Analysis reveals cut off score?

[DefenceDoom]

Hi everyone,

I have a research project where I am looking at the psychometric properties of a new questionnaire about body image.

I have two samples: an eating disorder sample and an undergrad control sample.

I wanted to see how well my questionnaire discriminates between clinical cases and controls. My committee member recommended using a discriminant function analysis (although maybe an ROC curve would have been more appropriate?)

My committee member said that the discriminant function analysis would help me to figure out my "cutoff score", or the score at which you can classify people into one group or the other.

However, I'm not sure where in the output/results to find this cutoff score? And I'm a bit too embarassed to ask him. I can't seem to find it anywhere.

Thanks very much

 

[the42up]

its the mean of the mean of both groups (if you are using equal sample, otherwise you need to weight the given means).

so you take the mean of the control group and the mean of your treatment, and then take the average of those two. That gives you your cut-off.

Good luck.

 

[DefenceDoom]

Thanks so much!!!