UNIB Statistics Lecturer Presents Small Area Estimation Model for Poverty Analysis in Bengkulu at MaGeStiC 2025 / Dosen Statistika UNIB Paparkan Model Small Area Estimation untuk Analisis Kemiskinan di Bengkulu pada MaGeStiC 2025

Jember, September 18, 2025 – Dr. Etis Sunandi, S.Si., M.Si., a lecturer from the Department of Mathematics, Faculty of Mathematics and Natural Sciences, University of Bengkulu, participated as a presenter at The 2nd International Conference on Mathematics, Geometry, Statistics, and Computation (MaGeStiC 2025) organized by the University of Jember. On this occasion, he presented his research titled “Bias Reduction using Small Area Estimation of the Beta-Binomial Model for Poverty Data Analysis in Bengkulu Province.”

The study was conducted in collaboration with a research team consisting of Sigit Nugroho, Dr. Idhia Sriliana, and Titin Siswantining (University of Indonesia), Laily Nissa Atul Mualifah (IPB University), as well as supervised students from the University of Bengkulu.

In his presentation, Dr. Etis Sunandi explained that the Small Area Estimation (SAE) model based on Beta-Binomial Adjusted Profile Hierarchical Likelihood (SAE-BB-APHL) was developed to address bias and overdispersion issues in poverty data at the sub-district level in Bengkulu Province. Using data from SUSENAS 2022 and PODES 2022, the model incorporates auxiliary variables such as the number of BPJS (national health insurance) participants, employment status, education level, and the number of health facilities to produce more accurate estimates.

The research findings indicate that the SAE-BB-APHL method yields smaller error values (MSE) compared to the direct estimation approach, thus providing more precise poverty mapping in areas with limited samples. The model has proven effective in supporting data-driven regional development planning.

This research is part of a Fundamental Research project funded by the Directorate of Research, Technology, and Community Service (DRTPM) of the Ministry of Education, Culture, Research, and Technology (Kemdikbudristek) under the DPPM Kemdiktisaintek scheme, with contract number 2856/UN30.15/PT/2025.

Dr. Etis Sunandi’s participation in MaGeStiC 2025 represents a tangible contribution from the Statistics Department of the University of Bengkulu in advancing quantitative methods for socio-economic issues, while also strengthening UNIB’s role in the international academic community in the fields of applied statistics and data science.

Through this participation, it is hoped that the research outcomes can be more widely applied in sectoral data analysis, particularly to support poverty mapping and data-based development policies in Bengkulu Province and other regions across Indonesia.

Jember, 18 September 2025 – Dosen Departemen Matematika, Fakultas MIPA Universitas Bengkulu, Dr. Etis Sunandi, S.Si., M.Si., berpartisipasi sebagai pemakalah dalam The 2nd International Conference on Mathematics, Geometry, Statistics, and Computation (MaGeStiC 2025) yang diselenggarakan oleh Universitas Jember. Pada kesempatan tersebut, beliau mempresentasikan hasil penelitian berjudul “Bias Reduction using Small Area Estimation of the Beta-Binomial Model for Poverty Data Analysis in Bengkulu Province.”

Penelitian ini dilakukan bersama tim kolaborator, yaitu Sigit Nugroho, Dr. Idhia Sriliana, Titin Siswantining (Universitas Indonesia), Laily Nissa Atul Mualifah (IPB University), serta mahasiswa bimbingan dari Universitas Bengkulu.

Dalam paparannya, Dr. Etis Sunandi menjelaskan bahwa model Small Area Estimation (SAE) berbasis Beta-Binomial Adjusted Profile Hierarchical Likelihood (SAE-BB-APHL) dikembangkan untuk mengatasi permasalahan bias dan overdispersi pada data kemiskinan tingkat kecamatan di Provinsi Bengkulu. Dengan menggunakan data SUSENAS 2022 dan PODES 2022, model ini memanfaatkan variabel bantu seperti jumlah peserta BPJS, status pekerjaan, pendidikan, dan jumlah fasilitas kesehatan untuk menghasilkan estimasi yang lebih akurat.

Hasil penelitian menunjukkan bahwa metode SAE-BB-APHL mampu menghasilkan nilai galat (MSE) yang lebih kecil dibandingkan metode estimasi langsung (direct estimation), sehingga memberikan hasil pemetaan kemiskinan yang lebih presisi di wilayah dengan sampel terbatas. Model ini dinilai efektif untuk mendukung perencanaan pembangunan berbasis data di daerah.

Penelitian ini merupakan bagian dari penelitian Fundamental Reguler yang didanai oleh Direktorat Riset, Teknologi, dan Pengabdian kepada Masyarakat (DRTPM) Kemdikbudristek melalui skema DPPM Kemdiktisaintek, dengan nomor kontrak 2856/UN30.15/PT/2025.

Kehadiran Dr. Etis Sunandi dalam MaGeStiC 2025 menjadi bentuk kontribusi nyata dosen Statistika Universitas Bengkulu dalam pengembangan metode kuantitatif untuk isu-isu sosial ekonomi, sekaligus memperkuat peran UNIB dalam komunitas akademik internasional di bidang statistika terapan dan data science.

Melalui partisipasi ini, diharapkan hasil penelitian dapat diterapkan lebih luas dalam analisis data sektoral, terutama untuk mendukung pemetaan kemiskinan dan kebijakan pembangunan berbasis data di Provinsi Bengkulu dan wilayah lainnya di Indonesia.