Page 1 of 1

Hair Jr. et al (2017) A primer on...edition 2e

Posted: Tue Feb 20, 2018 9:02 pm
by rmoquin1
I was wondering if anyone could confirm something for me.

On page 180 of “A Primer on Partial Least Squares Structural Equation Modeling edition 2e), Hair Jr. et al. (2017) indicate that formative measurement items csor_2, csor_4, qual_2, qual_3, and qual_4 are clearly below .001 (c.f. Exhibit 5.29 p.184).

Is this correct? Am I reading this incorrectly? Since the confidence level contains a negative value and the p-value is greater .001, doesn’t this necessitate that these items be removed from the model?

Thanks

Re: Hair Jr. et al (2017) A primer on...edition 2e

Posted: Wed Feb 21, 2018 3:18 pm
by cringle
They are not significant. But this does not mean you need to remove these items. If you read the book, you will find that you need to inspect the outer loadings next. Since they are above 0.5 you want to keep these items. In general, you may want to keep theoretically established items even though they are not significant in a certain context (in that example for German mobile phone provider data).

Regards
Christian