There has been a mistake in my data collection so that a group of my respondents in the first data collection T1 answered a questionnaire on a 6-point scale and in the second data collection T2 answered the same questionnaire on a 5 - point scale. To determine whether they answered differently because I gave them a different scale, I'd like to compute differential item functioning. Is this possible in this context?
Look at it this way: at point T2, the temporal effect on your construct of interest consists of two elements: (1) your intervention and (2) change in construct operationalization. Unfortunately, you have no way of teasing these two apart, unless you had a large enough number of participants answering a 5-point questionnaire at T1 -- or a 6-point questionnaire at T2.
Look at it this way: at point T2, the temporal effect on your construct of interest consists of two elements: (1) your intervention and (2) change in construct operationalization. Unfortunately, you have no way of teasing these two apart, unless you had a large enough number of participants answering a 5-point questionnaire at T1 -- or a 6-point questionnaire at T2.
thanks for your answer! I have a group answering on a 5-point scale at T1 but these are people from a different experimental group so also not comparable I guess.
The thing is I did calculate differential item functioning at T1 for these two groups (between-subjects data) and didn't have it. So I just wanted to know whether it's not possible to do the same for within-subject data?
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Applied Statistics
- Survey Design
- Questionnaire