Page 1 of 1

High Composite Reliability

Posted: Sat Aug 05, 2017 2:48 pm
by ankit
Dear Professor

I am using previously developed and validated models and now testing them through PLS SEM (SPLS 3). In my knowledge those models have been tested only once in the past.
Its a reflective measurement model and the problem occurred in stage 5a (i.e. Internal Consistency Reliability - CR of some constructs is above 0.90).

"Values above 0.90 (and definitely> 0.95) are not desirable because they indicate that all the indicator variables are measuring the same phenomenon and are therefore unlikely to be a valid measure of the construct."

Is there any way by which I can reduce the CR or use any other criterion like Cronbach's Alpha????

Please help me as I am at very last stage of my research. Time is an important constraint for me.

Re: High Composite Reliability

Posted: Sun Aug 06, 2017 1:53 pm
by jmbecker
Please search the forum. There have been discussions around this issue already.