Dear colleagues,
According to Wynne Chin (1998), the interpretation of formative indicators should be based on the weights. Like the canonical correlation analysis, the weights allow to understand the make-up of each variable. That is to say, these provide information about how each formative indicator contributes to the respective composite construct.
In this respect, I wonder what is the meaning of a negative weight for a formative indicator when its loading is positive:
(weights) CI
ci1 0,1607
ci2 -0,1176
ci3 0,4867
ci6 -0,0712
ci8 0,4657
(loadings) CI
ci1 0,6533
ci2 0,4512
ci3 0,8398
ci6 0,6101
ci8 0,8085
Thanks in advance for your help
José L. Roldán
Negative weight for a formative indicator
Negative weight for a formative indicator
José Luis Roldán, Ph.D.
Department of Business Administration and Marketing (University of Seville)
jlroldan@us.es
Department of Business Administration and Marketing (University of Seville)
jlroldan@us.es
- Diogenes
- PLS Super-Expert
- Posts: 899
- Joined: Sat Oct 15, 2005 5:13 pm
- Real name and title:
- Location: São Paulo - BRAZIL
- Contact:
Hi José,
I have tried some examples of mine and I have get results like yours when the formative indicators are correlated (collinearity).
Trying with 8 simulated indicators (non-correlated), one of them was negative (loading and weight) and the others are positive.
When I have made the sum of (weight * loading) = 1
but in your case the result was 0,79
Did you forgot some indicators ?
I hope this help.
Bido
I have tried some examples of mine and I have get results like yours when the formative indicators are correlated (collinearity).
Trying with 8 simulated indicators (non-correlated), one of them was negative (loading and weight) and the others are positive.
When I have made the sum of (weight * loading) = 1
but in your case the result was 0,79
Did you forgot some indicators ?
I hope this help.
Bido