PLS‐SEM: Prediction‐oriented solutions for HRD researchers

Abstract

Structural equation modeling, often referred to as SEM, is a well‐established, covariance‐based multivariate method used in Human Resource Development (HRD) quantitative research. In some research contexts, however, the rigorous assumptions associated with covariance‐based SEM (CB‐SEM) limit applications of the method. An emergent complementary SEM approach, partial least squares structural equation modeling (PLS‐SEM), is a variance‐based SEM method that provides valid solutions and overcomes several limitations associated with CB‐SEM. Despite PLS‐SEM's increasing popularity in many social sciences disciplines, the method has yet to gain traction in the field of HRD. An accessible overview of the method, including potential advantages for HRD research and extant methodological advancements, is provided in this article with the goal of encouraging productive dialogue in the field of HRD surrounding the PLS‐SEM approach. We present an emergent analytical tool for quantitative HRD research, offer practical guidelines for researchers to consider when selecting a SEM method, and clarify assessment stages and up‐to‐date evaluation criteria through an illustrative example.

A perspective on using partial least squares structural equation modelling in data articles

“…After the results computation, authors should report the criteria for reflective and formative measurement model assessment and the structural model. Several articles and textbooks provide overviews of which PLS-SEM results and criteria should be reported and how [69][70][71] . Table 1 provides a summary of key criteria as discussed in, for example, Hair, Hult, Ringle and Sarstedt [43] .…”

Social Undermining and Employee Creativity: The Mediating Role of Interpersonal Distrust and Knowledge Hiding

“…We performed a confirmatory factor analysis (CFA) to check the reliability and construct validity of the measurement scales. All the alpha coefficient, composite reliability, and average variance extracted (AVE) values exceeded the threshold of 0.7, 0.7, and 0.5, respectively [ 41 ]. For the determination of convergent validity, factor loading of the scale indicators on their corresponding factors was evaluated.…”

“…For the determination of convergent validity, factor loading of the scale indicators on their corresponding factors was evaluated. All the factor loadings of the scale indicators crossed the recommended value of >0.7 [ 41 ], showing the strong correlation with their respective constructs. The factor loading of only one item, EC10 (i.e., 0.667), was below the threshold level of >0.7.…”

“…This item was retained in the final model, since loadings >0.5 and <0.7 are considered acceptable if the composite reliability and AVE values of the respective construct meet the recommended criteria (CR > 0.7 and AVE > 0.5) [ 42 ]. To determine the discriminant validity, we used the Fornell–Larcker criterion, where the square root of the AVE of each construct should be higher than the construct’s respective correlation with all the other latent variables [ 41 ]. Table 1 and Table 2 depict the acceptable results of the measurement properties of all the scales.…”

The Mediating - Moderating Model of Interpersonal Conflict and Knowledge Hiding: Evidence from Advertising Agencies

“…Confirmatory factor analysis (CFA) was performed to evaluate the properties of the estimating scale. We computed alpha coefficients (> 0.7), average variance extracted (AVE >0.50), and composite reliability (CR >0.7) to evaluate the goodness of the measuring scale (Legate et al, 2021). For descriptive statistics (See Table 1).…”