IPMA analysis & non-significant structural path

[RasRina88]

Hi everyone,

I have two questions about IPMA analysis:
#1 - Should I test the structural model first before running the IPMA or can I just straight away run the IPMA?
#2 - If any structural path in the model shows non-significant result, say path A (predictor) to D (outcome) is not significant, must I include this predictor too in the analysis?

Looking forward to receiving your answer. Thank you.

[jmbecker]

You should first evaluate your measurement and structural model with the usual steps. It only makes sense to run IPMA on a valid model.

You usually do not delete insignificant paths. The idea of the analysis is to map importance and performance of your predictors. Non-significant predictors are not important, but that is also a finding within the larger analysis and you might want to contrast them against the important predictors.

[RasRina88]

Thank you Dr Becker. Your answer really helps me.