A primer on integrated generalized structured component analysis

Abstract

Purpose The purpose of this paper is to present integrated generalized structured component analysis (IGSCA) as a versatile approach for estimating models that contain both components and factors as statistical proxies for the constructs. The paper sets out to discuss the how-tos of using IGSCA by explaining how to specify, estimate, and evaluate different types of models. The paper’s overarching aim is to make business researchers aware of this promising structural equation modeling (SEM) method. Design/methodology/approach By merging works of literature from various fields of science, the paper provides an overview of the steps that are required to run IGSCA. Findings from conceptual, analytical and empirical articles are combined to derive concrete guidelines for IGSCA use. Finally, an empirical case study is used to illustrate the analysis steps with the GSCA Pro software. Findings Many of the principles and metrics known from partial least squares path modeling – the most prominent component-based SEM method – are also relevant in the context of IGSCA. However, there are differences in model specification, estimation and evaluation (e.g. assessment of overall model fit). Research limitations/implications Methodological developments associated with IGSCA are rapidly emerging. The metrics reported in this paper are useful for current applications, but researchers should follow the latest developments in the field. Originality/value To the best of the authors’ knowledge, this is the first paper to offer guidelines for IGSCA use and to illustrate the method's application by means of the GSCA Pro software. The recommendations and illustrations guide researchers who are seeking to conduct IGSCA studies in business research and practice.

Hotel Selection Attributes and Booking Intention: Challenges of Hotel Industry in the Post COVID-19 Pandemic Era

“…The standardized root mean residuals (SRMR) value for a properly model fit was zero or nearby (Hu & Bentler, 1999). The SRMR value in this study was 0.078 (N=157), indicating that the model fit was moderately acceptable (Hwang et al, 2023). R 2 values of 0.75 are considered substantial, 0.50 moderate and 0.25 weak (Hair et al, 2019).…”

Leadership Styles and Organisational Effectiveness in Selected Public Universities in Uganda

“…For adaptability, it was high (mean = 3.56), flexibility was moderate (mean = 3.42) and productivity was also moderate (mean = 3.34). The HTMT correlations in Table 2 measuring discriminant validity showed that all the values were below 0.90, which is the maximum, suggesting the existence of discriminant validity (Hwang et al, 2023). Therefore, the constructs independently measured the study variables.…”

Evaluation of consumer technology experience: The mediation role of customer satisfaction and its impact on customer loyalty and word of mouth.

“…For example, "the idea that PLS data may be used to validate a measurement model is a fabrication; the PLS path estimate cannot be utilized in null hypothesis significance testing." While PLS is "not a panacea," Hwang et al (2023) responded to these worries about this approach is "a substantial method that warrants a prominent place in any empirical researcher's statistical arsenal." Furthermore, even with complex sample sizes, PLS-SEM analysis yields statistically significant findings ) Sarstedt et al, 2022).…”