My samples are different in size and I need to see if the alpha coefficient in one group is significantly different from that in the other group. My data are in SPSS (I am not an R user, unfortunately). If there is some independent program for doing this, I would be interested to know as well.
Thank you.
I do not know of anyway to compare two alpha coefficients, but coefficient alpha is not dependent on sample size.
Alpha is composed of only two things, the average correlation and the number of items. So if you have the same number of items in each sample, then you need to find a way to compare the average correlations.
You may want to use Split-Half method of Internal consistency test of reliability.
Also you may want to use PPMC to have an idea of the correlation between the two samples and the direction of the correlation. This may give you insight to how related the samples are.
I do not need to know if the samples are related. I just would need to see if the two alphas are significantly different.
Hello Ranjana,
Leonard Feldt has been publishing on this topic (for both independent and dependent estimates of Cronbach's alpha) since 1969. Here's a freely available link that addresses comparing estimates from two independent samples (as well as other applications): https://files.eric.ed.gov/fulltext/ED291755.pdf
There are plenty of other references before and after this, if you're interested.
Good luck with your work.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS