Advanced SEM for SDGs: Assessing Data Integrity in Private Healthcare

Structural Equation Modeling is a widely used research methodology in various fields and SDGs. it is crucial to understand SEM’s core principles, assumptions, and criteria. In this study, SEM is employed to assess normality, handle missing data, and minimize sampling errors. Confirmatory Factor Analysis plays a central role in evaluating the model fit with the data by hypothesizing the number of latent factors and their corresponding indicators. First, the normality of the data is evaluated using Skewness and Kurtosis, which should range between -2 and +2 to meet SEM’s normality assumption. The study focuses on independent variables i.e., self-esteem, social interaction and social comparison, with perceived social support as a mediating variable and psychological well-being as the dependent variable. A sample of 400 respondents confirmed the data’s normal distribution. Secondly, missing data is addressed, with questionnaires containing over 30% missing information excluded to avoid bias. Lastly, sampling errors are minimized using the 10-times rule, which ensures that the sample size is more than ten times the maximum number of paths to any latent variable. The final SEM model are evaluated using fit indices, all indicate good model fit. When fit indices are unsatisfactory, re-specification, reviewing modification indices, and adjusting factor structures can improve model fit with alternative estimation methods. This research adds both theoretical and practical value by enhancing the understanding and application of SEM in SDGs’ healthcare context, while also offering actionable strategies for improving SEM model fit.