Peer Influence, Student Ethics, and Learning Independence as Predictors of Elementary Students' Academic Discipline in Science Education: Evidence from Indonesia

Abstract

Academic discipline is a foundational condition for classroom learning, yet elementary teachers in many school systems report a decline in students' self-regulated conduct, prompting renewed attention to the social and dispositional habits that may underlie it. This study tested whether peer influence, student ethics, and learning independence jointly and individually predict fourth-grade students' discipline in science (IPAS) learning. Using an ex post facto, associative quantitative design, data were collected via Likert-scale questionnaires from all 38 fourth-grade students (saturated sampling) at a public elementary school and analyzed with multiple linear regression in IBM SPSS Statistics 26. All three predictors significantly and positively predicted learning discipline: peer influence (B = 0.328, t = 2.412, p = .021, standardized beta = .280), student ethics (B = 0.413, t = 2.375, p = .023, standardized beta = .271), and learning independence (B = 0.451, t = 5.355, p < .001, standardized beta = .527). Learning independence was the dominant predictor by both criteria. Together, the three variables significantly predicted learning discipline, F(3, 34) = 29.186, p < .001, explaining 72.0% of its variance (R-squared = .720; adjusted R-squared = .696). Diagnostic testing confirmed normally distributed residuals, no multicollinearity (tolerance > .61, VIF < 1.7 for all predictors), and no heteroscedasticity. The findings suggest that cultivating elementary students' learning discipline benefits from attention to peer relationships, ethical character, and self-regulated learning skills, not only classroom rule enforcement, with implications for character-education policy in comparable elementary school systems.