Since 1977 MMR has be the choice method for analyzing interaction effects in the social sciences (Aguinis & Gottfredson, 2010). What problems might I expect to encounter when using categorical variables in a MMR? This relates to my dissertation. Thanks!
Ref:
Aguinis, H., & Gottfredson, R. K. (2010). Best-practice recommendations for estimating interaction effects using moderated multiple regression. Journal of Organizational Behavior, 31(6), 776-786.
Hi - can you provide more context? MMR can involve continuous or categorical moderators as IVs; categorical DVs would lead to multinomial logistic regression (w/ interactions given moderation). Good luck with your dissertation!
Thanks, Frederick!
My IVs and the DV are continuous, whilst the moderator (educational level) is categorical. I wanted to challenge myself so I chose to do this (MMR) instead of an hierarchical MR; however, I am now wondering if I should transform the MVs into interval measurements so that all of the variables are continuous. Is it less complicated to have all variables continuous, rather than having a mix?
hi Ruth - If educational level is scaled so that it is linear (e.g., every change in level corresponds to the same increase in education), then you could enter it as a continuous variable. If you aren't sure, and there are not many educational levels (say
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis