SEM is a multivariate technique that combines aspects of multiple regressions, and can estimate a series of inter-related dependence relationships simultaneously (Hair et al., 2009; Byrne, 2009). This technique can incorporate both unobserved variables (latent) and observed variables (manifest) in both a measurement model and a structural model. In a structural model, SEM provides the ability to measure the structural relationships between the set of unobserved variables while explaining the amount of unexpected variance (Byrne, 2009). As SEM depicts the structural relationships between variables, it is a model of relationships among constructs that takes a confirmatory approach to the analysis of structural theory relating to some phenomena. In SEM, the causal process is presented by a series of structural equations and, to enable a clearer conceptualization of the theory, the structural relations are modelled pictorially (Byrne, 2009).
76
SEM takes a confirmatory approach rather than an exploratory approach to data analysis, and can provide explicit estimates of error variance parameters. According to Hair et al. (2009) , SEM is the best multivariate procedure for testing both the construct validity and the theoretical relationships between a set of concepts represented by multiple measured variables. In addition, SEM is a powerful technique that combines measurement model and structural model into a simultaneous test (Hair et al., 2009; Aaker and Bagozzi, 1979). However, while a covariance-based SEM (CB-SEM) using analysis software such as AMOS, has been more popular in business research, a more recently dominant approach of SEM is the partial least square SEM (PLS-SEM) approach which, according to the latest work of Hair and his colleagues (2011), is more useful than CB-SEM. In this study, they make a comparison and rule of thumb for selecting CB-SEM or PLS-SEM.
Table 3.4: Rule of Thumb for CB-SEM or PLS-SEM Selection (Hair et al., 2011)
Research Goals
If the goal is predicting key target constructs or identifying key “driver” constructs, select PLS-SEM.
If the goal is theory testing, theory confirmation, or comparison of alternative theories, select CB-SEM.
If the research is exploratory or an extension of an existing structural theory, select PLS-SEM.
Measurement Model Specification
If formative constructs are part of the structural model, select PLS-SEM.
Note that formative measures can also be used with CB-SEM but to do so it requires accounting for relatively complex and limiting specification rules.
If error terms require additional specification, such as covariation, select CB-SEM. Structural Model
77 If the structural model is complex (many constructs and many indicators), select PLS-
SEM.
If the model is non-recursive, select CB-SEM. Data Characteristics and Algorithm
If your data meet the CB-SEM assumptions exactly, for example, with respect to the minimum sample size and the distributional assumptions, select CB-SEM. Otherwise, PLS-SEM is a good approximation of CB-SEM results.
Sample size considerations:
o If the sample size is relatively low, select PLS-SEM. With large data sets, CB- SEM and PLS-SEM results are similar, provided that a large number of indicator variables are used to measure the latent constructs (consistency at large).
o PLS-SEM minimum sample size should be equal to the larger of the following: (1) ten times the largest number of formative indicators used to measure one construct or (2) ten times the largest number of structural paths directed at a particular latent construct in the structural model.
If the data are to some extent non-normal, use PLS-SEM; otherwise, under normal data conditions, CB-SEM and PLS-SEM results are highly similar, with CB-SEM providing slightly more precise model estimates.
If CB-SEM requirements cannot be met (e.g., model specification, identification, nonconvergence, data distributional assumptions), use PLS-SEM as a good approximation of CB-SEM results.
CB-SEM and PLS-SEM results should be similar. If not, check the model specification to ensure that CB-SEM is appropriately applied. If not, PLS-SEM results are a good approximation of CB-SEM results.
Model Evaluation
If you need to use latent variable scores in subsequent analyses, PLS-SEM is the best approach.
If your research requires a global goodness-of-fit criterion, then CB-SEM is the preferred approach.
78
Following Hair et al.’s (2011) rule of thumb, a PLS-SEM approach is the most appropriate and effective method for the current research model based on the research objectives, complexity of model and sample size.
3.11 Chapter Summary
In summary, chapter three reviews the research design and research methodology adopted for this study. This is followed by a discussion about the definition of variables and the process of survey development. The pretesting of measures and a pilot test were explained, with a focus on reliability and validity test results. The unit of analysis, population and sample size, and data collection techniques were then specified. The chapter concluded with an explanation of data analysis strategies using PLS software.
79
CHAPTER FOUR – PRELIMINARY ANALYSIS AND SEM PROCEDURES
4.1 Introduction
This chapter describes the data preparation, reliability and validity assessment of measurements, and the techniques employed in this study to validate the integrative research model and test the hypotheses. First, the chapter describes the checking of the data, outliers and multivariate assumptions prior to commencing the statistical analysis. Second, it presents the assessment of reliability and validity through SPSS. Third, it describes the preliminary analysis of the Pearson correlations. Fourth, the issue of common-method variance in the present research is addressed, and finally, the rationale for data analysis approach selection is presented.