Hello,
im doing a research, where my entire population comprises 17 units ( firms) so basically im studying them all ( census) , however, im wondering how to re-test the reliability of my questionnaire given that im using pre-established measurement scales but that are used in different context ? Can i still use Cronbach’s alpha in this case (given the small size of the population ) ? Or are there other reliability tests that i may use ? Thank you
Cronbach's alpha does not directly consider the number of cases. Arguably, since you have population data, you don't even need to be concerned about a margin of error around the correlations between items. Therefore, if alpha is, in other respects, appropriate for assessing the reliability of your scales, the fact that you are using it for a (small) population should not matter.
Burke D. Grandjean thank you so much for your insights, i appreciate it.
I agree with Burke D. Grandjean about the sample size issue, but I am less certain on not being concerned about the "margin error around the correlations between items." Coefficient alpha is based on the average correlation among among the items, so if that is low, the sum or the items will have a large component of random error, which will limit the maximum correlation that the scale can have with other items.
- Badges
- Science topic
- Similar topics
- Medicine
- Epi
- Age Groups
- Children