Examining Data and Measurement Model Specification in SEM: An Illustration from Management Development

Abstract

Data analysis is highly critical for a value added research output but very
tricky to handle by the researchers. Each statistical technique in research
methodology has its own nuts and bolts that the researcher has to take care
of. The purpose of present study is to present the most important aspects,
issues and procedures to examine the characteristics of data and
relationships of interest prior to Structural Equation Modeling technique.
Through literature review the authors have noted some main issues and
procedures in examination of data prior to a SEM analysis. Major issues
discussed in the paper are model complexity, sample size, nature of data, and
measurement model fit. An example in the field of Management Development
(MD) is also presented to explain the procedure of data analysis in SEM.
Findings of the research revealed that by devoting considerable time and
effort on examining and exploring the nature of data and the relationships
among variables, before the application of this technique, can help
researchers in resolving procedural issues that eventually lead to better
prediction and reliability of results. The present study contributes to
literature on SEM by providing a more holistic view of data examination
before SEM analysis and practical guidance for researchers to use SEM
more effectively.