Uncovering and Treating Unobserved Heterogeneity with FIMIX-PLS: Which Model Selection Criterion Provides an Appropriate Number of Segments?

Sarstedt, Marko; Becker, Jan-Michael; Ringle, Christian M.; Schwaiger, Manfred

doi:10.1007/bf03396886

Cited by 219 publications

(130 citation statements)

References 53 publications

Supporting

Mentioning

129

Contrasting

Order By: Relevance

“…To overcome the problem of stopping in a local optimum solution, PLS-IRRS requires conducting several runs with random starting partitions per pre-defined number of segments to ensure obtaining a final solution that is at least close to the optimum segmentation solution (Becker et al, 2013;Sarstedt, Becker, Ringle, & Schwaiger, 2011). Similar concerns and recommendations apply for latent class segmentation (Wedel & Kamakura, 2000).…”

Section: Figure 1: the Pls-irrs Algorithmmentioning

confidence: 99%

“…Similar concerns and recommendations apply for latent class segmentation (Wedel & Kamakura, 2000). Sarstedt, Becker, et al (2011) show in their simulations on finite mixture partial least squares segmentation in PLS (FIMIX-PLS; Hahn, Johnson, Herrmann, & Huber, 2002) that 10 (20) runs return the optimum solution with a probability of 80 (90) percent. Since the PLS-IRRS algorithm is very fast, as discussed later, we recommend carrying out at least 10 runs to ensure ending at least close to optimum segmentation solution.…”

Section: Figure 1: the Pls-irrs Algorithmmentioning

confidence: 99%

“…One may run FIMIX-PLS (Hahn et al, 2002; to determine the number of segments and, then, run PLS-IRRS to obtain the final segmentation solution. Alternatively, the systematic selection of the most appropriate number of segments may involve PLS-IRRS runs with different numbers of pre-specified segments (for details on the model selection see Becker, Ringle, Sarstedt, and Völckner (2014) and Sarstedt, Becker, et al (2011). For these results, the objective function outcome (e.g., the weighted least squares' maximum average 2 w R value) is one possible strategy to decide on the best-suited number of segments when using PLS-IRRS.…”

Section: Figure 1: the Pls-irrs Algorithmmentioning

confidence: 99%

See 2 more Smart Citations

Segmentation of PLS path models by iterative reweighted regressions

Schlittgen

Ringle

Sarstedt

et al. 2016

Journal of Business Research

Self Cite

View full text Add to dashboard Cite

Uncovering unobserved heterogeneity is a requirement to obtain valid results when using the structural equation modeling (SEM) method with empirical data. Conventional segmentation methods usually fail in SEM since they account for the observations but not the latent variables and their relationships in the structural model. This research introduces a new segmentation approach to variance-based SEM. The iterative reweighted regressions segmentation method for PLS (PLS-IRRS) effectively identifies segments in data sets. In comparison with existing alternatives, PLS-IRRS is multiple times faster while delivering the same quality of results. We believe that PLS-IRRS has the potential to become one of the primary choices to address the critical issue of unobserved heterogeneity in PLS-SEM.

show abstract

Section: Figure 1: the Pls-irrs Algorithmmentioning

confidence: 99%

Section: Figure 1: the Pls-irrs Algorithmmentioning

confidence: 99%

Section: Figure 1: the Pls-irrs Algorithmmentioning

confidence: 99%

See 1 more Smart Citation

Segmentation of PLS path models by iterative reweighted regressions

Schlittgen

Ringle

Sarstedt

et al. 2016

Journal of Business Research

Self Cite

View full text Add to dashboard Cite

show abstract

“…Against this background, we first replicate R&E's setup by estimating a two-construct model with 100 observations, indicator loadings of 0.6, 0.7, and 0.8, and two conditions for the structural Note: We discarded all simulation studies that examine the performance of methodological extensions of the PLS method, for example, to capture unobserved heterogeneity (Becker, Rai, Ringle, & Vö lckner, 2013;Ringle, Sarstedt, & Schlittgen, 2014;Ringle, Sarstedt, Schlittgen, & Taylor, 2013;Sarstedt, Becker, Ringle, & Schwaiger, 2011) or assess moderating effects (Chin, Marcolin, & Newsted, 2003;Henseler & Chin, 2010 …”

mentioning

confidence: 99%

Common Beliefs and Reality About PLS

Henseler

Dijkstra

Sarstedt

et al. 2014

Organizational Research Methods

Self Cite

2,426

769

View full text Add to dashboard Cite

This article addresses Rö nkkö and Evermann's criticisms of the partial least squares (PLS) approach to structural equation modeling. We contend that the alleged shortcomings of PLS are not due to problems with the technique, but instead to three problems with Rönkkö and Evermann's study: (a) the adherence to the common factor model, (b) a very limited simulation designs, and (c) overstretched generalizations of their findings. Whereas Rö nkkö and Evermann claim to be dispelling myths about PLS, they have in reality created new myths that we, in turn, debunk. By examining their claims, our article contributes to reestablishing a constructive discussion of the PLS method and its properties. We show that PLS does offer advantages for exploratory research and that it is a viable estimator for composite factor models. This can pose an interesting alternative if the common factor model does not hold. Therefore, we can conclude that PLS should continue to be used as an important statistical tool for management and organizational research, as well as other social science disciplines.

show abstract

“…As soon as if there are more parameters, similar to this current research where we have analyzed the satisfaction with services in respect to three different tourisms (medical tourism, gastronomical tourism and business tourism) stances; the one question arises; does the parameter differs from group to group? In order two address that we have conducted PLS based multi group analysis (MGA) [29,30]. The Table 3 below has been illustrated in more specific mode in order to facilitate the understanding of the analysis output.…”

Section: Partical Least Based Multi Group Analysis (Pls-mga)mentioning

confidence: 99%