Multilevel Structural Equation Modeling (Multilevel SEM) Dengan Pendekatan Bayesian (Studi Kasus: Kejadian Gizi Buruk Di Provinsi Aceh)

Simamora, Kamaluddin (2026) Multilevel Structural Equation Modeling (Multilevel SEM) Dengan Pendekatan Bayesian (Studi Kasus: Kejadian Gizi Buruk Di Provinsi Aceh). Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6003232002-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (1MB) | Request a copy

Abstract

Masalah gizi buruk, seperti stunting, wasting, dan underweight, masih menjadi tantangan serius dalam bidang kesehatan masyarakat di Indonesia, termasuk di Provinsi Aceh. Berbagai faktor turut berkontribusi terhadap permasalahan gizi buruk. Penelitian ini bertujuan untuk menganalisis faktor-faktor yang mempengaruhi gizi buruk di Provinsi Aceh menggunakan pendekatan Multilevel Structural Equation Modeling (Multilevel SEM) yang memungkinkan pemodelan hubungan kausal antara variabel-variabel laten pada dua tingkat analisis, yaitu level individu dan level kabupaten secara simultan. Estimasi parameter dilakukan menggunakan pendekatan Bayesian melalui Markov Chain Monte Carlo (MCMC) dengan algoritma Gibbs Sampling. Hasil analisis menunjukkan bahwa data memiliki struktur hierarkis yang mendukung penerapan model multilevel, sebagaimana ditunjukkan oleh nilai Intraclass Correlation Coefficient (ICC) yang mengindikasikan variasi pada tingkat individu dan kabupaten/kota. Model pengukuran pada within dan between level menunjukkan bahwa sebagian besar indikator merepresentasikan konstruk laten secara memadai, dengan kualitas pengukuran yang lebih kuat pada level individu. Pada model struktural, pola asuh dan lingkungan berpengaruh signifikan terhadap gizi buruk pada tingkat individu, sementara hubungan pada level kabupaten/kota belum menunjukkan signifikansi yang kuat. Evaluasi model berbasis Bayesian menunjukkan konvergensi yang baik berdasarkan nilai Potential Scale Reduction (PSR), meskipun hasil Posterior Predictive Checking mengindikasikan bahwa model belum sepenuhnya mereplikasi struktur data observasi.
=================================================================================================================================================
Malnutrition, such as stunting, wasting, and underweight, remain serious challenges in the field of public health in Indonesia, including in Aceh Province. Various factors contribute to the persistence of malnutrition. This study aims to analyze the factors influencing malnutrition in Aceh Province using a Multilevel Structural Equation Modeling (Multilevel SEM) approach, which allows for the simultaneous modeling of causal relationships among latent variables at two levels of analysis, namely the individual level and the district/municipality level. Parameter estimation was conducted using a Bayesian approach through Markov Chain Monte Carlo (MCMC) with the Gibbs sampling algorithm. The results indicate that the data exhibit a hierarchical structure that supports the application of a multilevel model, as reflected by the Intraclass Correlation Coefficient (ICC), which suggests the presence of variation at both the individual and district/municipality levels. The measurement models at the within and between levels show that most indicators adequately represent the latent constructs, with stronger measurement quality observed at the individual level. In the structural model, parenting patterns and environment significantly influence malnutrition at the individual level, while the relationship at the district/city level has not yet shown strong significance. Overall model evaluation based on the Bayesian framework indicates satisfactory convergence according to the Potential Scale Reduction (PSR) values; however, the Posterior Predictive Checking results suggest that the model has not fully replicated the observed data structure.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Bayesian, Gibbs Sampling, Gizi Buruk, Multilevel, Stunting Bayesian, Gibbs Sampling, Malnutrition, Multilevel, Stunting
Subjects:	Q Science Q Science > QA Mathematics > QA279.5 Bayesian statistical decision theory.
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49101-(S2) Master Thesis
Depositing User:	Kamaluddin Simamora
Date Deposited:	29 Jan 2026 06:51
Last Modified:	29 Jan 2026 06:51
URI:	http://repository.its.ac.id/id/eprint/131020

Actions (login required)

View Item