Hey,
I have a model with a Formative second-order construct formed by two LCs, the first is reflective while the second is formative. My questions are :
1) Can a formative second-order construct be established with a reflective and formative LCs?
If Yes :
2) Should use repeated or two-stage approach?
3) Which Mode (A OR B) should be used to estimate the formative second-order construct?
Another question :
When I run Pls Algorithm for a formative construct with two indicators, the outer loading of one indicator is 1.00. Is that normal ? what does that mean?
Thank you
Second ordre construct
-
- SmartPLS Developer
- Posts: 1284
- Joined: Tue Mar 28, 2006 11:09 am
- Real name and title: Dr. Jan-Michael Becker
Re: Second ordre construct
1) Generally yes. I think it is not well covered in the literature, but it is possible and also conceptually plausible.
2+3) In such a situation I would prefer the two-stage approach, because you exactly have the problem that it is hard to decide which Mode to use at the higher-order constructs (i.e., Mode A or Mode B).
For the second-stage variable it would be straightforward as it is formative is should by default get Mode B (if not other reasons like extreme multi-collinearity suggest a different procedure).
Other question:
An outer loading of 1 is certainly unusual if you have two or more indicators. As the loadings are bivariate correlations between construct and indicator, it means that the construct is equal to the indicator. You have perfect correlation.
2+3) In such a situation I would prefer the two-stage approach, because you exactly have the problem that it is hard to decide which Mode to use at the higher-order constructs (i.e., Mode A or Mode B).
For the second-stage variable it would be straightforward as it is formative is should by default get Mode B (if not other reasons like extreme multi-collinearity suggest a different procedure).
Other question:
An outer loading of 1 is certainly unusual if you have two or more indicators. As the loadings are bivariate correlations between construct and indicator, it means that the construct is equal to the indicator. You have perfect correlation.
Dr. Jan-Michael Becker, BI Norwegian Business School, SmartPLS Developer
Researchgate: https://www.researchgate.net/profile/Jan_Michael_Becker
GoogleScholar: http://scholar.google.de/citations?user ... AAAJ&hl=de
Researchgate: https://www.researchgate.net/profile/Jan_Michael_Becker
GoogleScholar: http://scholar.google.de/citations?user ... AAAJ&hl=de
-
- PLS Junior User
- Posts: 1
- Joined: Wed Nov 28, 2018 7:04 pm
- Real name and title: Shweta doctoral student
Re: Second ordre construct
Could you pls clarify when using repeated indicator approach , did you treat all the indicators as reflective or formative? How did you arrive at that decision?
Just to elaborate I have two first order constructs which make the final order constructs. One of it is reflective (1st order) and other formative (1st order). How do I treat them when using repeated indicator approach (all reflective/formative).
Just to elaborate I have two first order constructs which make the final order constructs. One of it is reflective (1st order) and other formative (1st order). How do I treat them when using repeated indicator approach (all reflective/formative).
-
- SmartPLS Developer
- Posts: 1284
- Joined: Tue Mar 28, 2006 11:09 am
- Real name and title: Dr. Jan-Michael Becker
Re: Second ordre construct
Usually, you should attach them in the way the higher-order component is operationalized. If it is formative add all repeated indicators formative. If it is reflective add all repeated indicators reflective.
The latter case, howeher, would generally make not much sense in your example (e.g., mixed formative and reflective lower-order on a reflective higher-order component).
The latter case, howeher, would generally make not much sense in your example (e.g., mixed formative and reflective lower-order on a reflective higher-order component).
Dr. Jan-Michael Becker, BI Norwegian Business School, SmartPLS Developer
Researchgate: https://www.researchgate.net/profile/Jan_Michael_Becker
GoogleScholar: http://scholar.google.de/citations?user ... AAAJ&hl=de
Researchgate: https://www.researchgate.net/profile/Jan_Michael_Becker
GoogleScholar: http://scholar.google.de/citations?user ... AAAJ&hl=de