Hi,
There are two options in SmartPLS regarding the modeling - Formative and Reflective. I am keen to know the situations when these two are applied. If i can get a clarification on the same that will be of great help.
Best Regards,
Pritha
Relective or Formative
- Diogenes
- PLS Super-Expert
- Posts: 899
- Joined: Sat Oct 15, 2005 5:13 pm
- Real name and title:
- Location: São Paulo - BRAZIL
- Contact:
Re: Relective or Formative
Hi Pritha,
Reflective - we have a latent variable that is unobservable, than we use manifested variables as indicators (effect indicators). The LV is the cause and MVs are the effects. If the LV vary, all indicators will vary together. In this case we could use Cronbach´s Alpha and others internal consistency indexes.
Formative - The MVs are weighted to form de LV. It´s not necessary that the MVs have correlations between them. You could think LV as dependent variable and MVs as independent variables in a multiple regression.
For more information see
JARVIS, Cheryl B.; MACKENZIE, Scott B.; PODSAKOFFF, Philip M. A critical review of construct indicators and measurement model misspecification in marketing and consumer research. Journal of Consumer Research. v. 30, n.2, p. 199-218, 2003.
BOLLEN; Kenneth; LENNOX, Richard. Conventional wisdom on measurement: a structural equation perspective. Psychological Bulletin. v.110, n.2, p. 305-314, 1991.
CHIN, Wynne W. Partial Least Squares for researchers: na overview and presentation of recent advances using the PLS approach. USA: University of Houston, 2000.
http://www.misq.org/archivist/vol/no22/ ... mntry.html
Best regards
Reflective - we have a latent variable that is unobservable, than we use manifested variables as indicators (effect indicators). The LV is the cause and MVs are the effects. If the LV vary, all indicators will vary together. In this case we could use Cronbach´s Alpha and others internal consistency indexes.
Formative - The MVs are weighted to form de LV. It´s not necessary that the MVs have correlations between them. You could think LV as dependent variable and MVs as independent variables in a multiple regression.
For more information see
JARVIS, Cheryl B.; MACKENZIE, Scott B.; PODSAKOFFF, Philip M. A critical review of construct indicators and measurement model misspecification in marketing and consumer research. Journal of Consumer Research. v. 30, n.2, p. 199-218, 2003.
BOLLEN; Kenneth; LENNOX, Richard. Conventional wisdom on measurement: a structural equation perspective. Psychological Bulletin. v.110, n.2, p. 305-314, 1991.
CHIN, Wynne W. Partial Least Squares for researchers: na overview and presentation of recent advances using the PLS approach. USA: University of Houston, 2000.
http://www.misq.org/archivist/vol/no22/ ... mntry.html
Best regards
Prof. Dr. Diogenes de Souza Bido
-
- PLS Expert User
- Posts: 54
- Joined: Wed Oct 19, 2005 5:53 pm
- Real name and title:
Here's an illustrative example for both types of measurement:
Assume you would want to measure how drunk a particular person is.
To do it reflectively, you could:
- measure his/her blood alcohol level
- observe how well he/she can still walk in a straight line
- test his/her reaction time
- etc.
All these measures should covariate strongly as they describe effects of "drunkenness".
To do it formatively, you could:
- count the number of beers a person has had
- count the number of glasses of wine a person has had
- count the number of shots a person has had
- etc.
All these measures don't have to covariate (depending on the drinking habits of that person), but still cause "drunkenness".
Assume you would want to measure how drunk a particular person is.
To do it reflectively, you could:
- measure his/her blood alcohol level
- observe how well he/she can still walk in a straight line
- test his/her reaction time
- etc.
All these measures should covariate strongly as they describe effects of "drunkenness".
To do it formatively, you could:
- count the number of beers a person has had
- count the number of glasses of wine a person has had
- count the number of shots a person has had
- etc.
All these measures don't have to covariate (depending on the drinking habits of that person), but still cause "drunkenness".