Hi all,
for my dissertation I have a 2x2 mixed anova with manipulated IV, manipulated MED, continuous IV, continuous DV. I also have 3 continuous control variables.
How can I check easily whether to include my control variables or not? For example I did a Chi-squared test for my categorical variables, to see if they differ between-subjects. They did not differ so I excluded them from my analysis.
I did an ANCOVA first with my MED (perceived quality) as DV (afterwards will do an ANCOVA for my DV (purchase intention)).
However in this ANCOVA I cannot include my moderators (which are two variables: perceived product quality of brand 1 and perceived product quality of brand 2, not dummies, and I cannot make dummies out of them).
Does anyone have a simple suggestion to check whether I should include my continuous control variables or not? As I cannot add my moderator (which is not continuous but manipulated in the MED (AND DV).
(P.s. note: I am relatively new to this).
Thanks in advance!
Kind regards, Maayke
It sounds like you would be better off approaching this analysis as a linear mixed model (i.e., from a regression framework), rather than anova/ancova.
Also, I'd advise you not to decide which controls to include or exclude based on your chi-square approach, variable by variable. A particular control variable might only be significant when other variables are also controlled, especially if you expect moderation/interactions. So you should probably estimate the full model, guided by the theoretical and empirical literature, before deciding to exclude non-significant controls.
- Badges
- Science topic