Regression Discussion > Difference between Covariate and Between-subject Factors
Hi, one related question. If I have a sample of 3 females and 10 males and I am running a 2x2 repeated measures ANOVA, shall I put "sex" as a between-subject factor? There are so few females that I guess it will skew the results? (I know I should use non-parametric here, but there isn't any equivalent of 2x2 ANOVA...) Thanks Jeremy!
KC Tam
I doubt very much that a model with only 13 people will produce valid results. However, to answer your question generally: "covariates" are typically continuous variables (or interval/scale), while "between-subject factors" are typically categorical variables.
As for whether you should include them in your model, I can't answer that question for you, but if you have no interest in their effects, I probably would keep them out (especially since your sample is already so small, giving you very little power to detect effects).
Jeremy Taylor
I am using a mixed-design ANOVA in SPSS (i.e. Analyze --> General Linear Model --> Repeated Measures), but I am confused by "Covariates" and "Between-subject factors". What is the difference between the two? For instance, I have a group of subjects doing the test for 4 times and their scores are thus the within-subject dependent variables. But they also vary in sex (binary variable) and age (interval variable). I've tried to put "sex" into the "Covariates" box and the "between-subject factors" box, and they yielded very different results.
If I am not interested in the effects of sex and age, but it's a standard procedure to report them. Should I put them into covariates?
Thank you very much!