Simulated data with two latent classes showing different local models.
Usage
data(simdata)
Arguments
Format
A data frame of simulated data with 400 observations on the following 14 variables.
mv1
first manifest variable of the block Price Fairness
mv2
second manifest variable of the block Price Fairness
mv3
third manifest variable of the block Price Fairness
mv4
fourth manifest variable of the block Price Fairness
mv5
fifth manifest variable of the block Price Fairness
mv6
first manifest variable of the block Quality
mv7
second manifest variable of the block Quality
mv8
third manifest variable of the block Quality
mv9
fourth manifest variable of the block Quality
mv10
fifth manifest variable of the block Quality
mv11
first manifest variable of the block Customer Satisfaction
mv12
second manifest variable of the block Customer Satisfaction
mv13
third manifest variable of the block Customer Satisfaction
group
a numeric vector
Details
The postulated model overlaps the one used by Jedidi et al. (1997) and by
Esposito Vinzi et al. (2007) for their numerical examples.
It is composed of one latent endogenous variable, Customer Satisfaction,
and two latent exogenous variables, Price Fairness and Quality.
Each latent exogenous variable (Price Fairness and Quality) has five
manifest variables (reflective mode), and the latent endogenous variable
(Customer Satisfaction) is measured by three indicators (reflective mode). Two latent classes showing different local models are supposed to exist.
Each one is composed of 200 units. Thus, the data on the aggregate level
for each one of the numerical examples includes 400 units. The simulation scheme involves working with local models that are different
at both the measurement and the structural model levels.
In particular, the experimental sets of data consist of two latent classes with
the following characteristics: (a) Class 1 - price fairness seeking customers - characterized by a strong relationship
between Price Fairness and Customer Satisfaction (close to 0.9) and a
weak relationship between Quality and Customer Satisfaction (close to 0.1),
as well as by a weak correlation between the 3rd manifest variable
of the Price Fairness block (mv3) and the corresponding latent variable; (b) Class 2 - quality oriented customers - characterized by a strong relationship
between Quality and Customer Satisfaction (close to 0.1) and a weak
relationship between Price Fairness and Customer Satisfaction (close to 0.9),
as well as by a weak correlation between the 3rd manifest variable (mv8) of the Quality
block and the corresponding latent variable.
References
Esposito Vinzi, V., Ringle, C., Squillacciotti, S. and Trinchera, L. (2007)
Capturing and treating unobserved heterogeneity by response based segmentation in
PLS path modeling. A comparison of alternative methods by computational experiments.
Working paper, ESSEC Business School. Jedidi, K., Jagpal, S. and De Sarbo, W. (1997)
STEMM: A general finite mixture structural equation model.
Journal of Classification14, pp. 23-50. Trinchera, L. (2007)
Unobserved Heterogeneity in Structural Equation Models:
a new approach to latent class detection in PLS Path Modeling.
Ph.D. Thesis, University of Naples "Federico II", Naples, Italy.