I am interested in carrying out an ANCOVA with a categorical covariate. I have a continuous dependent variable (level of depression), two independent variables: one being level of bullying (none, bully, victim) and an independent variable that I consider to be a moderator (gender), and then a a dichotomous covariate (language: English or Dutch).
How can I run this is SPSS? Or, is there another analysis I can use?
Hello Steph,
If the categorical covariate has only two levels, then you could just enter it as a covariate with a (0,1) or (1,2) coding. Covariates only influence analysis results to the degree that: (a) scores on the covariate relate to scores on the DV; (b) comparison groups differ on the covariate. If neither condition is present, the covariate doesn't make any difference (except for drawing away 1 df from the error term).
There is literature to suggest that building the variable into the design as another factor (also referred to as a blocking variable) may work better if the correlation between the covariate and DV is low.
Good luck with your work!
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS