To give you an example: let's suppose I have 3 subscales A, B and C, each with 5 item, and I know

  1. The Cronbach's alpha of each subscale (let's suppose: 0.80 for A, 0.85 for B and 0.90 for C),

  2. The Cronbach's alpha of the global (15-item) scale (let's suppose: 0.90 for A+B+C=S).

If I assume that the three between-subscale correlations (AB, BC and AC) are equal, is there a formula to estimate such parameter?

Given that Cronbach's alpha derives from mean inter-item correlation (https://stats.stackexchange.com/a/339775/159259), then I think the question can be restated as:

Can we derive the between-subscale correlation (assuming it is constant across pairs) from the mean inter-item correlation of each subscale and of the global scale?



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.