2

I have one manipulated independent variable (degree of co-creation) based on three surveys "no" "low" and "high" degrees of co-creation.

Furthermore, I have 4 dependent variables (purchase intention, loyalty, word of mouth, willingness to pay).

I want to test whether the degree of co-creation (IV) has a direct effect on my dependent variables.

What should I do? Regression by means of dummy variables, or MANOVA? Or can it be an added value to do both?

Alexis
  • 29,850
Daan
  • 21
  • 1
    Ordinal IV in modeling/regression is always a problem and you have to choose either to treat it as interval or nominal. But: in your case you have only one IV, and it may solve the problem. Why not to see the IV as the DV and the 4 DVs as IVs, - and use ordinal regression? – ttnphns Dec 07 '13 at 11:00
  • See also https://stats.stackexchange.com/questions/190156/t-tests-manova-or-logistic-regression-how-to-compare-two-groups – kjetil b halvorsen Aug 14 '22 at 13:56

1 Answers1

0

You have one predictor and four response variables, it's problem of multivariate regression particularly, MANOVA.

If you have only one predictor, you can use MANOVA. For more than one predictor, MANCOVA is used. Both MANOVA and MANCOVA are multivariate regression techniques.

Multivariate regression is done in SPSS using the GLM-multivariate.option. Here is a nice tutorial on the same:

https://statistics.laerd.com/spss-tutorials/one-way-manova-using-spss-statistics.php#procedure

IBM support doc can also be useful:

http://www-01.ibm.com/support/docview.wss?uid=swg21476743

If you prefer using R, R package mvtnorm can be used for this purpose.

Dr Nisha Arora
  • 1,044
  • 1
  • 10
  • 25