Respondents need to mark all four choices in ranking order (4 most likely, 1 least likely)
Eg. When I get an idea, I will............
I will thought about it over and over again. 4
I try to reflect similar situation in the past 2
I will analyse and see the odds of it 3
I will act upon the idea. 1
in my opinion (assuming that I understood your question and your examples of items right) it does not make sense to compute conbrachs alpha in that case. you will need cronbachs alpha to get the reliability of a (reflexive measured) scale, e.g. you have a couple of items (each from 1 to 5) which all measure a similar phenomenon, for example extraversion. so you need some items which correlate, and not items like in your case which have to be rank-ordered by participants.
exactly :-)
in that suggested case one may also try t-test or analyses of variances for repeated measures, to check out it Responses differ over time.
I agree with Michael's second response. You have only one question so you cannot check "internal consistency" in the sense of whether all the items in a scale measure the same thing, which is what Cronbach's alpha does.
Thank you everyone for your answers. Is there any other methods to check the reliability of this type of question ?
paralleltest or retest (as Michael mentioned above). e.g.: https://www.uni.edu/chfasoa/reliabilityandvalidity.htm
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Applied Statistics
- Survey Design
- Questionnaire