I have a group of 16 students in my sample. 14 students evaluated themselves (anonymously) on 24 competencies prior to an intervention. Following the intervention, 12 students evaluated themselves (again, anonymously) on the same set of competencies. Because I can't compare each individual's pre and post score, I want to compare the means of each of the competency self-evaluation scores, pre and post. I can't figure out what to do.
I should also say I tried a paired t-test two sample for means and a two-sample t-test assuming equal variances. I got back very strange results so I'm unsure if either test was the right one to use.
Mam kindly provide more details like what scales you have used? How many outcome variables do you have?
you just do a comparison of means on spss. it gives you the mean before and after, and the confidence interval. you don't have a large enough sample to do anything else as its underpowered. If the CI range is both over or under 0, then you can claim some significance - at least a trend, with caveats re the small sample size. you can run this for the overall score (all 24 items) or individual items. Or, if you have have items grouped into different sets of competencies, then you can run means for each competence (i.e., if you ask quetions that might fall into different types of competencies (knowledge, skills, confidence, etc) then you can get a score for each competence. Hope that makes sense. you do this by selecting the items you want to include before running the calculation. Comparing means is also easy by hand if you are not using spss - Excel probably does it too but I don't know how to do that. Look up comparing means on spss on google. Lots of info there.
If it is interval/ratio level data with related samples, it should be a related samples t-test. If the DV is measured on an ordinal scale, I think it would be the Wilcoxon T-test. I can check my Bannon book or loan it to you if that would help.
- Badges
- Science topic