Within-subject factor or different DVs?

#1
Hi all,

Something I came across when helping a colleague out. I can conceptualize one 'grouping' of his data as a within-subject factor or as different dependent variables. He's looking at the attributes of brands (functional/traits/values) and, in a survey, asks for the importance of each in a buying decision for a certain product.

I could see these things as different variables or as a within-subject factor. I'm leaning towards the latter, because A) they're all measured on the same Likert scale and B) he wants to compare the means of each category. But what, generally speaking, determines whether to see something as a within-subject factor or as different DVs?
 

css

New Member
#2
Hi Junes,
I am not an expert, but hopefully, my reply might result stimulating and you come with a solution for your question.

I would consider these measures as different DVs. In my opinion, the attributes of the brands seem to refer to "objective" properties of these brands while the survey seems to be asking about the subjective importance of these attributes in a particular situation (a buying decision). Besides, till my limited knowledge, considering these different measures as different DVs does not preclude comparing them (e.g. by calculating confidence intervals).

In a more general sense, I feel that repeated measures should be exactly that, replicates of the same variable in the same conditions. For other situations, I feel it might be more appropriate to use MANOVAs (maybe this article might be helpful O'brien, R. G., & Kaiser, M. K. (1985). MANOVA method for analyzing repeated measures designs: an extensive primer. Psychological bulletin, 97(2), 316). However, as I said before, and I am not a statistics' expert and probably other users of this forum can provide you better suggestions.