You are on page 1of 31

Partial Least Squares - Structural Equation Modeling (PLS - SEM)

Using SmartPLS 3.0

Associate Professor Dr. S. Mostafa Rasoolimanesh


School of Hospitality, Tourism, and Events
Taylor’s University
11th June – 2nd July, 2020
Assessment of Moderator
• Many theories in management, psychology, and other
disciplines rely on moderating variables: those which affect
the strength or nature of the relationship between two other
variables. (Dawson, 2013)
• Interaction effect analysis for continuous and scale variable, as well
as categorical with two groups
• Multi group analysis (MGA) for categorical moderator with two or
more than two groups
Y = b0 + b1X + b2M + b3 XM + ɛ
Assessment of Moderator
Assessment of Interaction Effect

• In the literature related to PLS path modeling, approaches for the analysis of
interaction effects between variables have been presented so far.
• First, Chin et al. (1996, 2003) developed the so-called product indicator
approach.
• Second, Henseler and Fassott (2010) and Chin et al. (2003) suggested using a
two-stage approach under certain circumstances.
• Third, based on an initial proposal by Wold (1982), the inventor of PLS, a hybrid
approach can be constructed.
• Finally, an orthogonalizing approach suggested by Little et al. (2006) for
modeling interactions among latent variables to the effect that it can be used
with PLS path modeling. (Henseler & Chin, 2010, p.84)
Assessment of Moderator
Assessment of Interaction Effect
Product Indicator Approach
Assessment of Moderator
Assessment of Interaction Effect
Two- Stage Approach
Assessment of Moderator
Assessment of Interaction Effect
• As the Monte Carlo simulation revealed, both the two-stage and the hybrid
approach have a high level of statistical power compared with the
orthogonalizing and especially the product indicator approach.
• Only in the case of few indicators and few observations, the orthogonalizing
approach seems to be advantageous.
• Recognizing the easy use of the two-stage approach on the one hand, and
the lack of available software implementation of the hybrid approach on the
other hand, it appears recommendable to apply the two-stage approach to
assess the significance of an interaction effect (Henseler & Chin, 2010,
p.105)
Assessment of Moderator
Assessment of Interaction Effect

• Fordescription of estimated parameters of integration effects,


product indicator approach is better than two stage approach.
(Henseler & Chin, 2010)
• Forformative moderator, two stage approach is suitable to
analyze interaction effect. (Henseler & Chin, 2010)
Assessment of Moderator
Interpretation of Interaction Effect

Y = (b0 + b2M + ɛ) + (b1 + b3 M) X

• The equation has been rearranged into a different form, representing a


regression of Y on X having the constant as well as the slope of the
exogenous variable X depending on the level of the latent moderator
variable •(Jaccard & Turrisi, 2003, p. 17). This form provides intuitive appeal
to the interpretation of interaction effects: An increase in the moderator
variable •of M 1 SD implies a change of the effect of X on Y by b3 . For
instance, if • M is standardized and increased from 0 to 1, the slope of
changes from b1 to b1 + b3 . (Henseler & Chin, 2010)
Assessment of Moderator
Interpretation of Interaction Effect
Assessment of Moderator

• Chin et al. (2003) state that a low effect size f 2 does not necessarily imply that the underlying
moderator effect is negligible: “Even a small interaction effect can be meaningful under extreme
moderating conditions, if the resulting beta changes are meaningful, then it is important to take
these conditions into account” (Chin et al. 2003, p. 211).

• According to Cohen (1988), f2 is assessed as:


0.02 small
0.15 medium
0.35 large
Assessment of Moderator
Three way Interaction
Y = b0 + b1X + b2M1 + b3 M2 + b4 XM1 +
b5XM2 + b6M1 M2 + b7X M1 M2 + ɛ
Assessment of Moderator
Three way Interaction
(Dawson, 2013)
Assessment of Moderator
Three way Interaction
(Dawson, 2013 )
Multi Group Analysis
• Several studies explicitly broach the issue of group-specific effects
in their research questions, ignoring population heterogeneity –
when performing PLS path modeling on an aggregate data level –
can seriously bias the results and, thereby, yield inaccurate
management conclusions (Sarstedt, Schwaiger, & Ringle, 2009).
For example, cross-national or cross-cultural differences are
related to observed heterogeneity, and should be considered in
the research to avoid bias results (Sarstedt et al., 2011, p.197).
Multi Group Analysis
• When the moderator is already categorical in nature,
• For example, gender, race, socioeconomic class
• Or with a “dichotomized” metric scaled (continuous) variable
• Transformed into a categorical variable
• Divide the moderating variable into “high” and “low” categories
• If indicators have interpretable mean, the grouping value is “high” for all indicators above
the mean, and “low” for all indicators below the mean.
• If indicators have no interpretable mean, use “high” grouping values for upper third, and
“low” for bottom third.
• Can also group based on a median split.
Multi Group Analysis
Multi Group Analysis
• Three approaches to multigroup analysis have been proposed within a PLS
path modeling framework thus far.
o The first approach, introduced by Keil et al. (2000), involves estimating model
parameters for each group separately, and using the standard errors obtained from
bootstrapping as the input for a parametric test. This method is generally labeled the
parametric approach (Henseler, 2007).
o Chin (2003b) proposed and further described a distribution free data permutation test
(Chin & Dibbern, 2010; Dibbern & Chin, 2005), because the parametric approach’s
distributional assumptions do not fit PLS path modeling’s distribution-free character.
This test seeks to scale the observed differences between groups by comparing these
differences to those between groups randomly assembled from the data.
o Henseler (2007) proposed and described another nonparametric procedure, which
directly compares group-specific bootstrap estimates from each bootstrap sample (see
also Henseler et al., 2009). (Sarstedt et al., 2011, p. 199).
Multi Group Analysis
Multi Group Analysis
Invariance Assessment

• Testing differences
• in path coefficients
• across groups
requires that the latent variables are created in the same
way for all groups.
Invariance Assessment

• Metric invariance:
• All factor loadings for like items are invariant across groups.
• Partial metric invariance:
• Many factor loadings (50%) for like items are invariant across
groups.
• Partial metric invariance is often the highest stage of measurement
invariance that one can obtain.
• Partial metric invariance is most often sufficient for group
comparisons.
Invariance Assessment
• Common factor models are among the most frequently used methods for assessing measurement invariance in SEM; however, PLS-
SEM is a composite model with LV scores calculated based on a composite model algorithm (Henseler et al., in press). Henseler et al.
(in press) suggested the measurement invariance of composites (MICOM) method for PLS-SEM.
• MICOM is a three-step process involving configural invariance assessment and compositional invariance assessment (Henseler et al.,
in press). To assess configural invariance, the following criteria must be satisfied:
• (a) the associated indicators for each construct should be identical across both models;
• (b) the methods of treating the data (e.g., coding, reverse coding, the handling missing values and outlier, and
standardization) should be identical; and
• (c) the algorithms applied to the inner and outer models should be identical (Henseler et al., in press).

 In addition, to assess compositional invariance, the score of identical composites (i.e., LVs) should be identical across the two
groups. Therefore, MICOM is used to estimate the correlation between identical LV scores across two models and should be
similarly correlated (i.e., have a correlation of 1). A correlation value of 1 between same composites (LVs) in each models indicates
compositional invariance (Henseler et al., in press).

 Equality of Composite Mean Values and Variances; While using a multigroup analysis requires establishing configural and
compositional invariance, running analyses on the pooled data level necessitates establishing the equality of the composites’ mean
values and variances (Henseler et al., in press).
Invariance Assessment
Reporting Multi Group Analysis
Thank you for your attention
References
• Akter, S., D'Ambra, J., & Ray, P. (2011). Trustworthiness in mHealth information services: an assessment of a hierarchical
model with mediating and moderating effects using partial least squares (PLS). Journal of the American Society for
Information Science and Technology, 62(1), 100-116.
• Albers, S. (2010). PLS and success factor studies in marketing Handbook of Partial Least Squares (pp. 409-425): Springer.
• Bagozzi, R. P., & Fornell, C. (1982). Theoretical concepts, measurements, and meaning. A second generation of multivariate
analysis, 2(2), 5-23.
• Bagozzi, R. P., & Yi, Y. (1988). On the Evaluation of Structural Equation Models. Academy of Marketing Science, 16(1), 74-94.
• Becker, J.-M., Klein, K., & Wetzels, M. (2012). Hierarchical latent variable models in PLS-SEM: guidelines for using reflective-
formative type models. Long Range Planning, 45(5), 359-394.
• Chin, W., Marcolin, B., & Newsted, P. (1996). A partial least squares latent variable modeling approach for measuring
interaction effects: results from a Monte Carlo simulation study and voice mail emotion/adoption study.
• Chin, W. W. (1998). The partial least squares approach for structural equation modeling. In G. A. Marcoulides (Ed.), Modern
methods for business research. Methodology for business and management. Mahwah, NJ: Lawrence Erlbaum Associates
Publishers.
• Chin, W. W. (2010). How to Write Up and Report PLS Analyses. In V. E. Vinzi, W. W. Chin, J. Henseler & H. Wang (Eds.),
Handbook of Partial Least Squares. London, New York: Springer.
• Chin, W. W., & Newsted, P. R. (1999). Structural equation modeling analysis with small samples using partial least squares.
Statistical strategies for small sample research, 1(1), 307-341.
• Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences, Second Edition. Hillsdale, New Jersey: Lawrence Erlbaum
Associates, Inc.
• Dawson, J. F. (2013). Moderation in Management Research: What, Why, When, and How. Journal of Business and Psychology, 1-19.
• Diamantopoulos, A., Riefler, P., & Roth, K. P. (2008). Advancing formative measurement models. Journal of Business Research,
61(12), 1203-1218.
• Diamantopoulos, A., & Winklhofer, H. M. (2001). Index construction with formative indicators: an alternative to scale development.
Journal of Marketing Research, 38(2), 269-277.
• Dijkstra, T. K., & Henseler, J. (2012). Consistent and asymptotically normal PLS-estimators for linear structural equations.
• Dijkstra, T. K., & Schermelleh-Engel, K. (2013). Consistent partial least squares for nonlinear structural equation
models. Psychometrika, 1-20.
• Fornell, C., & Larcker, D. F. (1981). Evaluation Structural Equation Models with Unobsevable Variable and Measurement error.
Journal of Marketing Research, 18(1), 39-50.
• Gefen, D., Rigdon, E. E., & Straub, D. (2011). EDITOR’S COMMENTS: An Update and Extension to SEM Guidelines for Administrative
and Social Science Research. MIS Quarterly, 35(2), iii- xiv.
• Gefen, D., Straub, D. W., & Boudreau, M.-C. (2000). AND REGRESSION: GUIDELINES FOR RESEARCH PRACTICE.
• Gerbing, D. W., & Anderson, J. C. (1988). An updated paradigm for scale development incorporating unidimensionality and its
assessment. Journal of Marketing Research (JMR), 25(2).
• Hair, J. F., Hult, G. T. M., Ringle, C., & Sarstedt, M. (2013). A primer on partial least squares structural equation modeling (PLS-SEM):
SAGE Publications, Incorporated.
• Hair, J. F., Ringle, C. M., & Sarstedt, M. (2011). PLS-SEM: Indeed a Silver Bullet. Journal of Marketing Theory and Practice, 19(2), 139-
151.
• Hair, J. F., Sarstedt, M., Ringle, C. M., & Mena, J. A. (2012). An Assessment of the Use of Partial Least Squares Structural Equation
Modeling in Marketing Research. Journal of the Academy of Marketing Science, 40, 414-433.
• Hair Jr, J. F., Sarstedt, M., Hopkins, L., & Kuppelwieser, V. G. (2014). Partial least squares structural equation modeling (PLS-SEM): An
emerging tool in business research. European Business Review, 26(2), 106-121.Hayes, A. F., & Preacher, K. J. (2010). Quantifying and
testing indirect effects in simple mediation models when the constituent paths are nonlinear. Multivariate Behavioral Research,
45(4), 627-660.
• Hayes, A. F., & Scharkow, M. (2013). The Relative Trustworthiness of Inferential Tests of the Indirect Effect in Statistical Mediation
Analysis Does Method Really Matter? Psychological science, 24(10), 1918-1927.
• Hayes, M. H. (2009). Statistical digital signal processing and modeling: John Wiley & Sons.
• Heeler, R. M., & Ray, M. L. (1972). Measure validation in marketing. Journal of Marketing Research, 361-370.
• Henseler, J., & Chin, W. W. (2010). A comparison of approaches for the analysis of interaction effects between latent variables using
partial least squares path modeling. Structural Equation Modeling, 17(1), 82-109.
• Henseler, J., & Sarstedt, M. (2013). Goodness-of-fit indices for partial least squares path modeling. Computational Statistics, 28(2),
565-580.
• Henseler, J., Ringle, C., & Sarstedt, M. (in press) Testing measurement invariance of composites using partial least
squares. International Marketing Review.
• Hulland, J. (1999). Use of Partial Least Squares (PLS) in Strategic Management Research: A Review of Four Recent Studies. Strategin
Management Journal, 20, 195-204.
• Lohmöller, J.-B. (1989). Latent variable path modeling with partial least squares: Physica-Verlag
Heidelberg.
• MacKinnon, D. P., Fairchild, A. J., & Fritz, M. S. (2007). Mediation analysis. Annual review of psychology,
58, 593.
• MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., & Sheets, V. (2002). A comparison of
methods to test mediation and other intervening variable effects. Psychological methods, 7(1), 83.
• Nunnally, J. C. (1978). Psychomtietric theory (second edition) New York: McGraw-Hill.
• Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling strategies for assessing and
comparing indirect effects in multiple mediator models. Behavior research methods, 40(3), 879-891.
• Preacher, K. J., & Kelley, K. (2011). Effect size measures for mediation models: quantitative strategies for communicating indirect
effects. Psychological methods, 16(2), 93.
• Real, J. C., Roldán, J. L., & Leal, A. (2012). From entrepreneurial orientation and learning orientation to business performance:
analysing the mediating role of organizational learning and the moderating effects of organizational size. British Journal of
Management.
• Reinartz, W., Haenlein, M., & Henseler, J. (2009). An empirical comparison of the efficacy of covariance-based and variance-based
SEM. Intern. J. of Research in Marketing, 26, 332–344.
• Riou, J., Guyon, H., & Falissard, B. (2015). An introduction to the partial least squares approach to structural equation modelling: a
method for exploratory psychiatric research. International journal of methods in psychiatric research.
• Ringle, C. M., Sarstedt, M., & Straub, D. W. (2012). Editor's comments: a critical look at the use of PLS-SEM in MIS quarterly. MIS
Quarterly, 36(1), iii-xiv.
• Rucker, D. D., Preacher, K. J., Tormala, Z. L., & Petty, R. E. (2011). Mediation analysis in social psychology: Current practices and new
recommendations. Social and Personality Psychology Compass, 5(6), 359-371.
• Sarstedt, M., Henseler, J., & Ringle, C. M. (2011). Multigroup analysis in partial least squares (PLS) path modeling: alternative
methods and empirical results. Advances in International Marketing, 22, 195-218.
• Tenenhaus, M., Vinzi, V. E., Chatelin, Y.-M., & Lauro, C. (2005). PLS path modeling. Computational statistics & data analysis, 48(1),
159-205.
• Urbach, N., & Ahlemann, F. (2010). Structural equation modeling in information systems research using partial least squares. Journal
of Information Technology Theory and Application, 11(2), 5-40.
• Wetzels, M., Odekerken-Schröder, G., & Oppen, C. v. (2009). Using Pls Path Modeling for Assessing Hierarchical Construct Models:
Guidelines and Empirical Illustration. MIS Quarterly 33(1), 177-195.

You might also like