Work discipline, competence, and job satisfaction predicting work achievement in an Indonesian agricultural research institute

Abstract

This study examined the associations between work discipline, competence, and job satisfaction and self-reported work achievement, distinguishing their joint explanatory variance from individual unique contributions. The entire workforce of the Bakrie Agriculture Research Institute in Kisaran, North Sumatra (N = 58) completed five-point Likert composites for work discipline (4–20), competence (5–25), job satisfaction (7–35), and work achievement (5–25). Ordinary least squares regression yielded F(3, 54) = 110.43, p < .001, R² = .860, and model f² = 6.14. All three conditional associations were positive and significant: competence, B = 0.390, 95% CI [0.171, 0.609], β = .378, f² = 0.237; job satisfaction, B = 0.240, CI [0.047, 0.433], β = .312, f² = 0.115; work discipline, B = 0.287, CI [0.006, 0.568], β = .280, f² = 0.078. Collinearity was substantial (VIF 4.316–7.187). The squared semi-partial correlations summed to .060, leaving approximately .800 as shared explanatory variance. Thus, substantial joint fit coexisted with comparatively small unique contributions. Competence had the numerically largest standardised coefficient and unique contribution, but this ordering was exploratory; no formal between-predictor comparison was reported. A sensitivity analysis indicated an 80% power threshold of f² = .140, which contextualises precision without establishing predictor superiority. The findings support positive associations and fuller reporting of uncertainty in small-census research, without causal attribution or a confirmed ranking of predictors.