Mean interitem correlations PreviousNext
Mplus Discussion > Categorical Data Modeling >
 JJL posted on Tuesday, March 19, 2013 - 4:24 pm

Perhaps I am missing something in the output, but I am trying to obtain the mean interitem correlation for a 12-item scale of categorical variables. Could you please help me with syntax? Also, at which version (e.g., Mplus 5.0) did this begin to be an option?

Thank you very much.

 Bengt O. Muthen posted on Tuesday, March 19, 2013 - 6:26 pm
What is the mean interim correlation?
 JJL posted on Tuesday, March 19, 2013 - 7:07 pm
Perhaps others refer to it as the "Average Interitem correlation?" Essentially, I have composite scales that I have created, so I need to make sure there is some degree of unidimensionality and a moderate level of agreement between the items (i.e., coefficient ranging from .20-.50 per Clark & Watson, 1995).

Since these items are categorical, I believe they deserve tetrachoric correlations if I am not mistaken (as opposed to Spearman's, e.g.).
 Linda K. Muthen posted on Wednesday, March 20, 2013 - 10:27 am
It sounds like you are referring to Cronbach's Alpha. See papers by Tenko Raykov.
 JJL posted on Wednesday, March 20, 2013 - 12:29 pm
I guess I had heard that Chronbach's alpha is a bad measure of unidimensionality when categorical variables are involved...but perhaps I am mistaken?
 Linda K. Muthen posted on Wednesday, March 20, 2013 - 12:39 pm
We don't endorse Cronbach's alpha in any case. Tenko has written on this and may have an opinion related to categorical outcomes. Neither what you mention nor Cronbach's alpha is an option in Mplus.
Back to top
Add Your Message Here
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Options: Enable HTML code in message
Automatically activate URLs in message