background: 100 students were given an unrelated text and a related text each. they then had to answer questions on each text, say how interested they were in each text, and then evaluate how much they liked a character in the text.
Please help!!!!! I am struggling with the statistics from the papers I need to review for my course! Thank you so much, everyone.
A repeated measures MANOVA showed a significant effect, F (3,47)=3149 (Fcrit 6.4), p
I'm not entirely sure what you are asking. Are you asking us to evaluate whether the provided interpretations are correct? Whether the statistical modeling approach itself is correct?
In any case, since the outcome scores are all Likert scale (or something similar - they all appear to be categorical), an ANOVA-based approach is inappropriate. ANOVA based approaches (whether it be a RM-ANOVA, a MANOVA, or whatever) require continuous outcome variables. So, right from the start I wouldn't trust the validity of any of the results of this model. That doesn't necessarily mean the results are wrong, but I would want to verify them with a more principled statistical approach. To see why this is the case, look at the results from the first univariate test. You say that the mean of the scores in the related group is 6.89 out of a possible total 7; however, you say the standard deviation of that mean is 1.18. You should IMMEDIATELY realize the problem with those numbers; 6.89+1.18>7. Your outcome measure can't be higher than 7 by design, but the ANOVA model is saying that there is >50% probability of an observed observation being higher than 7!
Thank you so much for your response, Ryan. I am indeed asking about both the modal and my interpretation, because I'm not doing well at all on either side.
I had been so caught up on understanding what f statistics are and so on that I didn't even notice the standard deviation problem. I must go back and check I copied the scores properly. But seeing the other problems I noted with the paper... Yes, I must decide if I want to bin this paper and look at another one without such issues, or go ahead with it anyway...
Thank you Ryan! I am very grateful.
I can’t see what the „between subjects factor“ is. As I understood it correctly, you have a within repeated measures factor „condition (related vs. Unrelated), so all participants read both stories. And you have 3 dependent variables „understanding“, „interest“, and „evaluation“. So, what is he between subjects factor?
Regarding the F-test: to put it simple, it is just the ratio of model variance and error variance. The higher the model variance in relation to the error, the larger the F-value.
- Badges
- Science topic