I am writing up the results from a experiment where I have a multi-level IV and am measuring impact on motivational (H1-4) and performance variables (H5-6). I am wondering if I should put all of the dependent variables together in the same ANOVA or separate them into a motivation-oriented and performance-oriented ANOVA. Any advice will help, but coupling your response with management or psychology journal citations is preferred. Thank you.
If all scores are from the same Ss, first do a a full ANOVA. Okay to separate if you had an a priori set of hypotheses. But most importantly plot scores to see what the variables for individual variables look like. When you do this, make sure that your scales are comparable.
If the two main variables are related, you need to take this into account. Without seeing the design, hard to say whether my advice stands. Whatever, you do need to take into account the fact that you had (at least) two experimental variables. If you thought they are related, treating them as independent is a no-no. Also consider using a co-variate.
What do you mean by "if I should put all of the dependent variables together in the same "? A mean variable from all your DV H-16 or a mean variable for H1-4 and a second for H5-6? For the latter idea, you would have to run 2 separate ANOVAs. But if this was your initial idea, I would refer to Michael's answer to use two MANOVAs and discriminant analyses as follow up analyses.
But have a look at Tabachnick and Fidell (2014) Using Multivariate Statistics, 6th ed.
Thank you for the response Rainer. I did not intend to suggest averaging the DVs, but entering them in the ANOVA at the same time. Said differently, should all 6 variables be entered in the same ANOVA, should the variables be separated into two ANOVAs with motivational variables in one and performance variables in the other, or does it not matter.
Perhaps some clarity for everyone is that the motivation variables consist of engagement and its three sub factors (vigor, dedication, and absorption). The performance variables are simply two different measures of performance. Thank you all for your input. Very helpful.
Dear Daniele,
that's why I asked about the averaging. An ANOVA naturally has only one DV, and one or more IVs, so how can you incorporate several DVs? The only possibility I can think of is to create a compound variable of the DVs (e.g. averaging) and use an ANOVA or using a MANOVA, where you can have two or more DVs (technically a compound variable is calculated within the MANOVA, but this is another thing).
Please share with us, what you intended to do, for me this is not so clear.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA