Journal of Statistics and Data Science

APPLICATION OF STRUCTURAL EQUATION MODELING (SEM) IN ANALYZING ACADEMIC PERFORMANCE

Wed, 31 Dec 2025 00:00:00 +0000

Academic performance is one of the ways to determine whether one has had a good education. There are multiple factors that can influence someone’s academic performance. The multivariate model in Structural Equation Modeling (SEM) is used to determine the relations between level of stress, level of resilience, and academic performance. The data for this study is collected using google form, distributed to active students of Prasetiya Mulya. The SEM model in this study is a combination of multivariate regression analysis and confirmatory factor analysis called structural regression model. There are three hypotheses made in this study. The modelled hypotheses will then be evaluated using a chi-square test, Root Means Squared Error of Approximation (RMSEA), Tucker Lewis Index (TLI), Akaike Information Criterion (AIC), and Bayesian Information Criterion (BIC). The chi square test shows that two of the three models are significant with the first model being the best one out of the three. The first model, which states that resilience and stress affect students’ academic performance, but stress and resilience have no correlation with one another, shows minimal discrepancy between the data and model estimation, shown by the chi square p-value of 0.454, RMSEA of 0.005, and TLI of 1.

A Diagnostic Approach Using Statistical Quality Control Tools for Root Cause Identification of Ceramic Glaze Defects

Alfred Ayo Ayenigba, O. Esther Taiwo, Hakeem Makanjuola OYEDIRAN — Tue, 30 Dec 2025 00:00:00 +0000

This study applies statistical quality control (SQC) tools to address glaze defects and strengthen process stability in the production of artisan ceramic mugs at Artisan Mugs Inc., with particular focus on the “Midnight Blue” product line. A Pareto analysis of 300 defective units identified glaze imperfections as the most frequent issue (54%), followed by handle cracks (21.7%). Root cause analysis using a fishbone diagram revealed multiple contributing factors spanning materials (e.g., elevated glaze viscosity), manpower (e.g., insufficient operator training), methods, machinery, measurement practices, and environmental conditions. To evaluate process stability, and R control charts were constructed from 25 production batches. Both charts demonstrated statistical control with no evidence of assignable causes, further validated by runs tests (). Process capability analysis indicated adequate but improvable performance, with indices of and for means, and and for ranges. These findings confirm that the process is stable, yet improvements are needed in centering and reducing variability.

Factors Affecting The Open Unemployment Rate in West Sumatera Province Using The Multivariate Adaptive Regression Spline Method

Amesha Putri Pratama, Devni Prima Sari — Thu, 30 Oct 2025 00:00:00 +0000

The Open Unemployment Rate (OUR) is the percentage value of the ratio of the number of open unemployed to the total labor force. During the 2017-2022 period, the OUR faced fluctuations that showed significant numbers that could have a negative impact on economic growth and community welfare. The purpose of this study is to determine the factors that affect OUR in West Sumatra Province in 2017-2022 using the Multivariate Adaptive Regression Splines (MARS) method. This research is an applied research with data in the form of secondary data obtained at the Central Bureau of Statistics (CBS) of West Sumatra Province. The analysis carried out produces the best model at the minimum Generalized Cross Validation (GCV) value of 1.31675 from a combination of BF = 28, MI = 2 and MO = 1 so that the factors that influence the percentage of OUR are gross regional domestic product at current prices, population, population aged 15 years and over who are in school, population aged 15 years and over who work with the main job, average monthly per capita expenditure on food, average monthly per capita expenditure on non-food and monthly minimum wage with R² adj of 81.6 percent.

Analysis of Effective Policy Strategies to Address Multidimensional Stunting Using the Biplot Method in Bengkulu Province

Tue, 30 Dec 2025 00:00:00 +0000

This study aims to analyze the factors contributing to stunting in Bengkulu Province. The method used is biplot analysis, by reducing the Principal Component Analysis (PCA) dimensions into two components. A quantitative approach was employed, involving ten variables representing health, nutrition, education, housing, food security, and social protection factors. The results indicate that Bengkulu City has characteristics that are significantly different from other regencies. The key contributing factors include limited access to basic health services (particularly the availability of skilled birth attendants and immunization coverage), high levels of food insecurity, low access to proper sanitation and safe drinking water, limited practice of exclusive breastfeeding, and low utilization of government assistance programs such as BPJS Kesehatan (National Health Insurance) and KPS/KKS (Social Welfare Cards). It is expected that the findings of this study can provide valuable insights and contribute to efforts in reducing the prevalence of stunting in Bengkulu Province.

Application of Principal Component Analysis (PCA) in Determining the Dominant Factors Affecting Women’s Interest in Entrepreneurship in the Lower Market of Bukittinggi

Nafisatuzzahara SY -, Devni Prima Sari — Mon, 29 Dec 2025 00:00:00 +0000

The success of women in traditional markets is often constrained by various interrelated factors that directly affect the local economy. This study aims to identify the dominant factors influencing women's interest in entrepreneurship at Pasar Bawah, Bukittinggi, using the Principal Component Analysis (PCA) method. Primary data were collected through questionnaires distributed to women entrepreneurs, covering variables such as job choice, entrepreneurial interest, self-empowerment, social environment, and risk tolerance. PCA was applied to reduce correlated variables into fewer uncorrelated principal components. The analysis resulted in three principal components, with the first component selected as the dominant factor due to its highest explained variance. This component, with an eigenvalue of 4.73, explains 47.35% of the total variance and includes variables such as interest in entrepreneurship, willingness to take risks, feeling empowered and useful, and high self-confidence. These findings highlight the importance of psychological and personal factors in women's entrepreneurial interest. The study suggests that government policies should focus on inclusive support such as access to microcredit, digital entrepreneurship training, and promotion of local products to improve the competitiveness of traditional markets and empower women entrepreneurs.