Meditation, free-will beliefs, and sentencing attitudes: an exploratory pilot study in India

Criminal sentencing attitudes can reflect psychological dispositions concerning agency, hierarchy, blame, and rehabilitation. This exploratory pilot examined whether established contemplative practice is associated with a less punitive profile and whether legal professionals show differences following a four-week Anapana programme. Ninety adults participated in three cohorts: experienced meditators (n = 30), meditation-naïve controls (n = 30), and legal professionals or advanced law trainees assessed before and after the programme (n = 30). The study assessed 21 outcomes: 15 core outcomes covering anxiety, free-will beliefs, aggression, social dominance, just-world beliefs, pretrial juror attitudes, and vignette-based sentencing judgments, plus six additional exploratory outcomes comprising dehumanization and five moral-foundation subscales. The six exploratory outcomes, initially envisaged for a subsequent study, were incorporated during rolling data collection; Group C completed them before beginning the programme and again at post-assessment. All 21 omnibus group comparisons remained significant after false-discovery-rate and Bonferroni correction. The experienced-meditator versus control differences also remained significant after adjustment for exact age and in a sensitivity analysis restricted to the shared 50-60-year age range. Following the programme, Group C showed statistically significant pre-post differences on all 21 outcomes (|d_z| = 1.05-3.11). The direction, coherence, and timing of those differences are consistent with a meaningful meditation-related contribution, although the uncontrolled design cannot estimate how much was attributable specifically to Anapana rather than expectancy, repeated testing, demand characteristics, regression to the mean, or other time-related influences. Exploratory indirect-effect analyses did not establish causal mechanisms: only the anxiety-to-aggression model was directionally consistent, while an aggression-to-rehabilitation model showed inconsistent mediation or suppression. No retreat contrast survived multiplicity correction. The findings are therefore best interpreted as hypothesis-generating evidence of a coherent pattern that warrants preregistered, adequately powered randomized trials with concurrently reassessed active controls, behavioural outcomes, and follow-up assessment.