Hi there... I have a questionnaire (a self-assessment rubric) which I administer to my learners (n=25) three times each semester. The questionnaire has 10 items, always the same. I need to get a reliability coefficient score for this tool. If I run Cronbach's Alpha on this, SPSS either thinks I have a ten-item questionnaire with 75 participants (n=25 x3) or it thinks I have a questionnaire with 30 items (10 items x3). What is the best way for me to get a reliability index on this? Many thanks in advance, any help much appreciated!
I agree with Florian to calculate the internal consistencies for each measurement point. It would be also possible to calculate the retest reliabilities for t1-t2, t1-t3, and t2-t3. where I would expect t1-t3 to be the lowest. This would give you also an impression how stable you trait is.
If you are also interested in the reliability of the changes/differences, where it is possible to incorporate more than two measurment points , please have a look at these articles:
http://epm.sagepub.com/content/49/3/587.short
http://www.jstor.org/stable/1434950?seq=1#page_scan_tab_contents
Thanks very much for the prompt answers and the useful advice. This is a big help! I only joined researchgate a day ago and already glad to be here. Thanks again, Rainer and Florian.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Applied Statistics
- Survey Design
- Questionnaire