Abstract

Student satisfaction is a critical factor in evaluating the effectiveness of educational institutions. Understanding which factors most strongly influence satisfaction can help administrators improve teaching quality, learning environments, and institutional services. While surveys often capture multiple features (such as teaching methods, course structure, resources, and peer interaction), it remains unclear which features contribute most significantly to overall satisfaction. Without systematic analysis, decision-making may rely on assumptions rather than data-driven insights. In this study, linear regression was employed to assess the relationship between multiple independent variables (survey-based features/questions) and the dependent variable (student satisfaction level). The dataset was preprocessed by handling missing values, normalizing inputs, and encoding categorical features. The regression model was then trained to estimate the contribution (regression coefficients) of each feature. Statistical significance tests were conducted to identify which predictors have the strongest influence. The model revealed that instructional quality, availability of learning resources, and timely feedback from faculty were the most significant factors influencing student satisfaction. Less impactful variables included extracurricular activities and campus facilities. The findings provide actionable insights for institutional decision-makers to prioritize resources toward factors with the highest impact on satisfaction.

Authors

M. Viju Prakash
British University Vietnam, Vietnam

Keywords

Student Satisfaction, Linear Regression, Feature Analysis, Educational Data Mining, Predictive Modeling