I have presented a Table in a paper where I have modeled data using Multiple Regression Analysis with multiple fixed effects, repeated measures and a random effect using PROC MIXED in SAS. One of the reviewers comments was that I should describe how much variation the models explain - either partial eta squared for each term or r squared for the model as a whole. From the output I have the covariance parameter estimates which gives me the variance of the random effect and the mean square error for the model. Would these two parameters be acceptable for describing the variation? Alternatively, how could I go about generating partial eta or R2 values?
@Jos Feys I understand that R2 values explain the proportion of variation that a model can explain however, I'm not sure how to obtain R2 values using PROC Mixed in SAS. I believe it requires manual calculation and I'm not sure how to go about calculating it. I would assume I'd need an adjusted R2 value as well
I think you are wright:
http://animsci.agrenv.mcgill.ca/StatisticalMethodsII/drvpseudor.pdf
https://communities.sas.com/t5/Statistical-Procedures/R-Square-for-Individual-Fixed-Effect-Estimates-in-Proc-Mixed/td-p/374684
https://www.ars.usda.gov/ARSUserFiles/3122/Kramer2005.pdf
Use proc reg:
https://blogs.sas.com/content/iml/2017/01/09/ods-output-any-statistic.html
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Publication