Scale validity testing (concurrent)

[ljsimkcd]

Hello

I need to test the validity of a scale we have created. We have another pre-validated scale to compare with for a concurrent validity test.

We are asking people to complete both question sets and then intend to correlate their responses but am not sure how exactly to go about this. We tested the scale items using cronbach's alpha with item-total correlation and alpha 'if item deleted'.

For the validity test do I correlate people's mean scores on both scales or do I correlate individual responses - I don't know sorry.

The other complicating fact is that there are different numbers of questions in the two scales - so people completing both scales (within the same questionnaire) will be answering slightly more questions on the scale we want to test than questions from the pre-validated scale. Does this matter?

Hope I have explained clearly?

Thank you for your help.

 

[wandatheavenger]

Assuming that you've gotten rid of any items that you've deemed poorly functioning through item analysis on your new scale, go ahead and calculate either sum-scores or mean-scores and correlate those together. The different number of items isn't a problem at all (it actually might make your scale a more reliable measure).