2011
DOI: 10.3917/grhu.080.0045
|View full text |Cite
|
Sign up to set email alerts
|

Les avantages et les limites de la méthode « Partial Least Square » (PLS) : une illustration empirique dans le domaine de la GRH

Abstract: Les méthodes d’équations structurelles (MES) sont aujourd’hui largement employées dans la recherche quantitative en GRH, lorsqu’il s’agit de tester des modèles de causalité complexes, incorporant plusieurs variables latentes. La procédure habituelle d’estimation repose sur des techniques d’analyse des relations de covariance entre les variables, mises en application dans les logiciels couramment utilisés dans les traitements de données issues des études empiriques (ex : Lisrel, Amos, EQS). L’objectif de cet ar… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
18
0
4

Year Published

2018
2018
2023
2023

Publication Types

Select...
5
2
1

Relationship

0
8

Authors

Journals

citations
Cited by 32 publications
(22 citation statements)
references
References 64 publications
0
18
0
4
Order By: Relevance
“…Contrary to the classical structural equation modelling (i.e. Lisrel method, M+), the PLS-method is based on variance analysis and not on covariance analysis which allows calculations on smaller samples [27][28][29][30][31].…”
Section: Discussionmentioning
confidence: 99%
“…Contrary to the classical structural equation modelling (i.e. Lisrel method, M+), the PLS-method is based on variance analysis and not on covariance analysis which allows calculations on smaller samples [27][28][29][30][31].…”
Section: Discussionmentioning
confidence: 99%
“…Structural equation modeling was originally developed to examine the multiple causal relationships. Then, their use gradually extended to analyze the validity of latent constructs and to conduct multi-group analysis [29].…”
Section: 3methodsmentioning
confidence: 99%
“…Moreover, according to [31,32], the PLS-SEM method is recommended when the number of subjects included in the sample size is small, the data are not normally distributed, and the objective of the study is to identify key drive variables of a model. A model of structural equations traditionally consists of two parts: the measurement model (made up of all the relationships between the indicators and variables or latent constructs that they contribute to their measurement) and the structural model, which deals with the set of relationships between constructs and represents the network of causal relationships that the researcher wishes to establish [29]. Figure 3 displays the way to conduct a PLS-SEM analysis according to [33].…”
Section: 3methodsmentioning
confidence: 99%
“…We opted for structural equation modeling (SEM) to statistically analyze our model, the interest of the latter lies essentially in its capacity to test simultaneously the existence of causal relationships between several latent variables. For this reason, we chose the modeling by Partial Last Squares (PLS) approach since it is more suitable in case of obtaining a correct prediction of the level of the independent variables according to the dependent variables, as well as it is well adapted to exploratory type analysis in which the estimation can be carried out on small samples (Lacroux, 2011). In general, this approach is adapted to predictive causal analyses with a situation of high complexity and low theoretical information (Fernandes, 2012).…”
Section: The Statistical Analysis Methods Usedmentioning
confidence: 99%