Discriminant validity assessment in PLS-SEM: A comprehensive composite-based approach

The article below puts forth a comprehensive composite-based perspective on how one can conduct discriminant validity assessment, in the context of structural equation modeling via partial least squares (PLS-SEM).

Rasoolimanesh, S. M. (2022). Discriminant validity assessment in PLS-SEM: A comprehensive composite-based approach. Data Analysis Perspectives Journal, 3(2), 1-8.

Link to full-text file for this and other DAPJ articles:

https://scriptwarp.com/dapj/#Published_Articles

Abstract:

This paper aims to discuss modern approaches to assess discriminant validity in the context of structural equation modeling via partial least squares (PLS-SEM). It illustrates the application of these approaches using the WarpPLS 7.0 software. The Fornell-Larcker criterion, crossloadings method, heterotrait-monotrait (HTMT) ratio, and full collinearity test have been discussed in this paper. A step-by-step guide is provided to assess discriminant validity using these four tests in WarpPLS 7.0. The first three criteria are applicable for reflective constructs, while the full collinearity test can be applied for both reflective and formative constructs. In different social science disciplines, a combination of reflective and formative constructs is a common practice, therefore reporting the full collinearity test for the assessment of discriminant validity can be an advantage.

Best regards to all!

Reliability assessment in SEM models with composites and factors: A modern perspective

The article below puts forth a modern perspective on how one can conduct reliability assessments in models with composites and factors, in the context of structural equation modeling via partial least squares (PLS-SEM).

Canatay, A., Emegwa, T., Lybolt, L. M. & Loch, K. D. (2022). Reliability assessment in SEM models with composites and factors: A modern perspective. Data Analysis Perspectives Journal, 3(1), 1-6.

Link to full-text file for this and other DAPJ articles:

https://scriptwarp.com/dapj/#Published_Articles

Abstract:

This paper’s focus is on reliability tests for both composite-based and factor-based analysis algorithms in structural equation modeling through partial least squares (PLS-SEM). We illustrate this analysis employing a widely used PLS-SEM software tool, WarpPLS. The results show the magnitude of differences between the two approaches, suggesting that the estimates of coefficients obtained using the factor-based approach are more conservative than those obtained using the corresponding composite-based approach.

Best regards to all!

Moderated mediation, segmentation delta method, J-curve emergence, and causality assessment: Article

The article below discusses moderated mediation and the related emergence of J-curve relationships, in a context that is relevant to researchers employing structural equation modeling via partial least squares (PLS-SEM).

The article lays out three steps to combine moderation and J-curve analyses, with the goal of more fully understanding the underlying moderated mediation relationships. It proposes a new segmentation delta method to test for J-curve emergence, as part of this framework.

Finally, the article discusses three causality assessment indices that are used to show that the model used in the article is generally sound in terms of causality.

Kock, N. (2021). Moderated mediation and J-curve emergence in path models: An information systems research perspective. Journal of Systems and Information Technology, 23(3), 303-321.

Link to full-text file for this article:

Click for PDF file

Abstract (structured):

Purpose. J-curve relationship analyses can provide valuable insights to information systems (IS) researchers. We discuss moderated mediation in IS research and the related emergence of J-curve relationships. Design/methodology/approach. Building on an illustrative study in the field of IS, we lay out three steps to combine moderation and J-curve analyses, with the goal of more fully understanding the underlying moderated mediation relationships. We propose a new segmentation delta method to test for J-curve emergence, as part of this framework. Findings. We show, in the context of this study, the complementarity of moderation and J-curve analyses. Research limitations/implications. Currently, IS researchers rarely conduct moderation and J-curve analyses in a complementary way, even though there are software tools, and related methods, which allow them to do so in a relatively straightforward way. Originality/value. Our analyses were conducted with the software WarpPLS, a widely used tool that allows for moderated mediation and J-curve analyses, in a way that is fully compatible with the set of steps presented in this paper.

Best regards to all!