Have you ever entered "objective" data into a PLS model, e.g. sales/profits growth, R&D intensity growth etc.?
My intention was to rely on that data gathered from databases for the dependent variables and have perceptual constructs as independent variables to avoid a single informant and common source bias. In that case, the dependent would have to be measured as single-item construct. Or is it even feasible to construct a latent variable from different objective measures, e.g. "economic success" made up of sales growth, ROI growth, market share growth?
Constructs and "objective" measures in PLS
The classic example of a construct with formative indicators is Socio-Economic-Status, which is very analogous to your Economic Success construct.
So yes, creating a latent variable from multiple "objective" measures is done.
As an aside, two points: 1) Even "objective" measures such as sales or profit growth can be susceptible to measurement error and so may be more objective than some other measures but no measure is completely objective; 2) Since PLS does not partial out measurement variance from true score variance, the point would seem to be largely immaterial anyway.
So yes, creating a latent variable from multiple "objective" measures is done.
As an aside, two points: 1) Even "objective" measures such as sales or profit growth can be susceptible to measurement error and so may be more objective than some other measures but no measure is completely objective; 2) Since PLS does not partial out measurement variance from true score variance, the point would seem to be largely immaterial anyway.
John J. Sailors, PhD
Associate Professor of Marketing
The University of St. Thomas
Opus College of Business
Minneapolis, MN
Associate Professor of Marketing
The University of St. Thomas
Opus College of Business
Minneapolis, MN
-
joerghenseler
- PLS Expert User
- Posts: 39
- Joined: Fri Oct 14, 2005 9:59 am
- Real name and title:
GIFI
Michel Tenenhaus (from HEC) presented a model similar to yours on the PLS'03 conference in Lisbon within a PLS tutorial.
It was based on GIFI data as far as I can remember.
He estimated political stability of a country. He used objective, and partially dichotomous variables like e.g. "whether a country is governed by a dictatorship"
Maybe you can ask him.
His URL is: http://www.hec.fr/hec/eng/professeurs_r ... hp?num=110)
It was based on GIFI data as far as I can remember.
He estimated political stability of a country. He used objective, and partially dichotomous variables like e.g. "whether a country is governed by a dictatorship"
Maybe you can ask him.
His URL is: http://www.hec.fr/hec/eng/professeurs_r ... hp?num=110)