THE 'BROKEN LINK' IN ORGANIZATIONAL TRAINING ROI: A STUDY OF BEHAVIOURAL CHANGE AND BUSINESS IMPACT IN INDIAN MAHARATNA COMPANIES

Abstract

India's Maharatna Central Public Sector Enterprises (CPSEs) training ecosystem observe a consistent accountability gap, in spite of the substantial investment in human resource development (HRD). This empirical study observe and evaluate the impact of training in seven selected Maharatna companies such as - IOCL, NTPC, ONGC, SAIL, GAIL, HPCL, and POWERGRID, (anonymously M1 to M7) using the Kirkpatrick's Framework of Four-Level Evaluation, namely Reaction, Learning, Behaviour, and Results. Using a quantitative and longitudinal research design, data collected from a stratified sample of 2,520 employees across Junior, Middle, and Senior management levels of these CPSE. Three standardized training interventions were observed—Supervisory Development, Digital Literacy Enhancement, and Effective Communication— where the evaluation was done using pre- and post-training assessments of the participating employees, dual-source of behavioural ratings, and HOD-rated business impact instruments. Statistical analyses for this study included Paired t-tests, One-Way ANOVA, Pearson Correlation, and Exploratory Factor Analysis. These tools were applied to test five primary research hypotheses. Findings confirm that training is highly effective at the knowledge level (Cohen's d = 3.37) but at the same time it reveals a statistically insignificant correlation (r = 0.08, p > 0.05) between supervisor-rated behavioural change stage (Level 3) and HOD-rated business results pr impact (Level 4)—labelled as the 'Broken Link.' Moreover, a 'Senior Learning Slump' was also observed, where the Senior Managers showed the highest overconfidence gap (4.2%) between their self-perception and their supervisor ratings. Diagnostic factor analysis unveiled Training Infrastructure as the dominant systemic variable—the poor facilities produced an effect of 8.5% drop in Overall Training Effectiveness (OTE), while the trainer quality variance was found to be statistically negligible for consideration. The study also concludes with actionable policy recommendations for CPSEs and Maharatna L&D practitioners, emphasizing a shift from activity-based to impact-based training evaluation as the future practice.