A Social Network Model of Direct Versus Indirect Reciprocity in a Corrections-Based Therapeutic Community

Theory Therapeutic communities (TC) for substance abuse depend on the mutual aid among a group of residents to positively support the recovery process. An example of this aid arises from the expectation that TC residents will provide affirmations to each other for actions that are considered positive toward recovery or beneficial to the resident community. TCs are effective at treating substance abuse, however, it is unclear how the cooperative behaviors among residents are maintained. In light of a recent call for TC process studies, this paper reports on a novel agent-based computer model in which affirmations are recorded as directional arcs in a social network. Reception of affirmations by agents can be modeled based on either direct reciprocity (DR) or indirect reciprocity (IR). Respectively, an agent focuses its reciprocal actions towards those from whom it has received affirmations, or an agent focuses its reciprocal actions towards other agents that have a reputation for sending affirmations. The resulting modeled social networks are then compared to the social networks found in two Ohio TCs. This study finds that IR more closely mimics the level of reciprocity and transitivity found in the TC network than does the DR model. The findings of this exploratory study suggest the mechanisms underlying DR are unlikely to explain the interactions among TC residents. Also of interest is that social network tools may have a useful place in the day-to-day operation of TCs.