Hi all, Is it alright to us the χ2 test for a measure variable over time? E.g., a true-false question which is measured in two rounds of a survey ( not necessarily repeated participants) Thanks for your help.
Hi. If the participants aren't repeated, then a chi-square test of association [or, independence] is fine. In this case, time is no different than any other factor, like location.
But you should be careful " not necessarily repeated participants " implies that some may be repeated and some may be not. This may overestimate an effect of independence (and I do not know a test for "partially repeated sample"). In case of repeated measures, an McNemar test may be appropriate.
Though we use Chi square test to determine association of dependent and independent variables in research, this test is not able to evaluate level of significance as compared to SPSS and STATA.
- Badges
- Science topic
- Similar topics
- Statistics
- Survey