BackgroundSupervisors play an essential role in implementation by diffusing and synthesizing information, selling implementation, and translating top management’s project plans to frontline workers. Theory and emerging evidence suggest that through these roles, supervisors shape implementation climate—i.e., the degree to which innovations are expected, supported, and rewarded. However, it is unclear exactly how supervisors carry out each of these roles in ways that contribute to implementation climate—this represents a gap in the understanding of the causal mechanisms that link supervisors’ behavior with implementation climate. This study examined how supervisors’ performance of each of these roles influences three core implementation climate domains (expectations, supports, and rewards).Materials and methodsA sequenced behavioral health screening, assessment, and referral intervention was implemented within a county-based child welfare agency. We conducted 6 focus groups with supervisors and frontline workers from implementing work units 6 months post-implementation (n = 51) and 1 year later (n = 40) (12 groups total). Participants were asked about implementation determinants, including supervision and implementation context. We audio-recorded, transcribed, and analyzed focus groups using an open coding process during which the importance of the supervisors’ roles emerged as a major theme. We further analyzed this code using concepts and definitions related to middle managers’ roles and implementation climate.ResultsIn this work setting, supervisors (1) diffused information about the intervention proactively, and in response to workers’ questions, (2) synthesized information by tailoring it to workers’ individual needs, (3) translated top managements’ project plans into day-to-day tasks through close monitoring and reminders, and (4) justified implementation. All four of these roles appeared to shape the implementation climate by conveying strong expectations for implementation. Three roles (diffusing, synthesizing, and mediating) influenced climate by supporting workers during implementation. Only one role (diffusing) influenced climate by conveying rewards.ConclusionsSupervisors shaped implementation climate by carrying out four roles (diffusing, synthesizing, mediating, and selling). Findings suggest that the interaction of these roles convey expectations and support for implementation (two implementation climate domains). Our study advances the causal theory explaining how supervisors’ behavior shapes the implementation climate, which can inform implementation practice.