Partial Least Squares Regression and Structural Equation Models: 2016 Edition

by G. David Garson

A graduate-level introduction and illustrated tutorial on partial least squares (PLS). PLS may be used in the context of variance-based structural equation modeling, in contrast to the usual covariance-based structural equation modeling, or in the context of implementing regression models. PLS is largely a nonparametric approach to modeling, not assuming normal distributions in the data, often recommended when the focus of research is prediction rather than hypothesis testing, when sample size is not large, or in the presence of noisy data.

Why we think the new edition is important:

Over 50% more coverage than the 2012 edition.
Many more illustrations and figures
Covers assessing model fit for reflective and formative models in greater depth.
Specialized topics like use of interaction terms, multigroup analysis, importance-performance matrix analysis, and the PRESS statistic.

Below is the unformatted table of contents.

PARTIAL LEAST SQUARES: REGRESSION AND STRUCTURAL EQUATION MODELS
Overview 8
Data 9
Key Concepts and Terms 10
Background 10
Models 12
Overview 12
PLS-regression vs. PLS-SEM models 12
PLS-DA models 13
Mixed methods 13
Bootstrap estimates of significance 13
Reflective vs. formative models 14
Confirmatory vs. exploratory models 16
Inner (structural) model vs. outer (measurement) model 16
Endogenous vs. exogenous latent variables 17
Mediating variables 17
Moderating variables 18
Interaction terms 20
Partitioning direct, indirect, and total effects 23
Variables 24
Case identifier variable 24
Measured factors and covariates 24
Modeled factors and response variables 24
Single-item measures 26
Measurement level of variables 26
Parameter estimates 27
Cross-validation and goodness-of-fit 27
PRESS and optimal number of dimensions 28
PLS-SEM in SPSS, SAS, and Stata 29
Overview 29
PLS-SEM in SmartPLS 29
Overview 29
PLS-SEM structural equation modeling with SmartPLS 29
Creating a PLS project and importing data 29
Validating the data 33
Creating the path model in SmartPLS 34
Reflective vs. formative models 37
Hiding the measurement model 38
Estimation options in SmartPLS 38
PLS algorithm 38
Blindfolding 39
Bootstrapping 40
Finite mixture PLS 41
Running the path model in SmartPLS 41
Data metric for centered data 42
Weighting scheme 43
The stopping (abort) criterion 43
Options 44
Saving the model 45
Other default settings 45
SmartPLS Output 45
Model report 45
Checking for convergence 46
Path coefficients 47
Bootstrapped significance 49
Assessing model fit 54
Latent variable crossloadings 66
Measurement (outer model) weights 67
Structural model path coefficients (inner model coefficients) 69
Latent variable correlations 69
Factor scores 70
Multigroup Analysis (MGA) with FIMIX-PLS 70
Overview: Unobserved heterogeneity 70
Comparing models with differing numbers of segments 71
Entropy 73
Path coefficients 74
T-tests of differences in path coefficients 75
Labeling the segments 75
Alternatives to PLS-FIMIX 76
Parametric MGA 77
Overview 77
PLS regression modeling with SmartPLS 78
PLS regression: SmartPLS vs SPSS or SAS 78
PLS regression: SPSS vs.

301 pages, Kindle Edition

First published May 15, 2012