A soccer-based intervention improves incarcerated individuals' behaviour and public acceptance through group bonding

Abstract

As incarceration rates rise globally, the need to reduce re-offending grows increasingly urgent. We investigate whether positive group bonds can improve behaviours among incarcerated people via a unique soccer-based prison intervention, the Twinning Project. We analyse effects of participation compared to a control group (study 1, n = 676, n = 1,874 control cases) and longitudinal patterns of social cohesion underlying these effects (study 2, n = 388) in the United Kingdom. We also explore desistance from crime after release (study 3, n = 249) in the United Kingdom and the United States. As law-abiding behaviour also requires a supportive receiving community, we assessed factors influencing willingness to employ formerly incarcerated people in online samples in the United Kingdom and the United States (studies 4-9, n = 1,797). Results indicate that social bonding relates to both improved behaviour within prison and increased willingness of receiving communities to support re-integration efforts. Harnessing the power of group identities both within prison and receiving communities can help to address the global incarceration crisis.

Fig. 1. Average number adjudications received per 100 people in prison in the 2 months after the programme.

Shown are the control group (top) and the treatment group (bottom). Visual representation of the control group (the intercept) and treatment group coefficient (intercept – estimated treatment effect, see intent-to-treat model 1 in Table A4 in Supplementary Information Section A).

Fig. 2. Social bonding with criminals both directly and indirectly (via perceived chances to desist) contributes to procriminal attitudes.

In the mediation diagram a, b, c and c′ are path coefficients representing unstandardized regression weights and robust standard errors (in parentheses). The c path coefficient represents the total effect of fusion to a criminal group on procriminal attitudes. The c′ path coefficient refers to the direct effect of criminal group fusion on procriminal attitudes. [95% CI] and P values: a = [−0.294, −0.039], P = 0.011; b = [−1.574, −0.152], P = 0.018; c = [0.667, 1.721], P < 0.001; and c′ = [0.509, 1.592], P < 0.001 (Table C4 in Supplementary Information Section C).