likert items questionnaire

[dulude]

Dear all,

I would like to ask for some advice regarding the data of a questionnaire i sent out. The questionnaire comprises 10 categegories. Each category consists of 8 likert items. 5 of these items are called enabler and the remaining 3 are performance items.

Now, I would like to analyze the impact of the enabler on the performance items. Thus, I want to identify the enabler with a significant influence. All items are 7 point likert items.

Does anyone know which analysis would be most appropriate for this purpose? Is it any kind of factional anova, or rather a stepwise regression?

Thank you very much.

 

[Karabiner]

Thus, I want to identify the enabler with a significant influence.

So you sum up the 3 performance items within each category and
perform 5 Spearman rank correlations, between the sum score
and each enabler.

With kind regards

K.

 

[dulude]

Thank you very much for your help. I appreciate it.

Your advice is seems very promising. However, with correlations I might not be able to argue for causalities of enablers on performance. Do you know by any chance, which method I could use to proof that?

 

[Injektilo]

Well, you can assess the impact they would have on performance, but not the causality. You'd need to set up a causal methodology to properly assess that.

As far as measuring the impact though, this is a matter of debate. Statistical purists would severely frown down on this because they consider Likert scales to be ordinal rather than interval data, but I would just run a regression. The problem is though that there is likely to be a strong amount of multicollinearity between your IVs. One solution to this is using Shapley Values Regression, which is a little trick commonly used in market research. If you can use R, check out the relaimpo package which allows to run this type of regression.