摘要：Traditional methods of process evaluation encompass what components were delivered, but rarely uncover how practitioners position themselves and act relative to an intervention being tested. This could be crucial for expanding our understanding of implementation and its contribution to intervention effectiveness. We undertook a narrative analysis of in-depth, unstructured field diaries kept by nine community development practitioners for two years. The practitioners were responsible for implementing a multi-component, preventive, community-level intervention for mothers of new babies in eight communities, as part of a cluster randomised community intervention trial. We constructed a narrative typology of approaches to practice, drawing on the phenomenology of Alfred Schutz and Max Weber's Ideal Type theory. Five types of practice emerged, from a highly 'technology-based' type that was faithful to intervention specifications, through to a 'romantic' type that held relationships to be central to daily operations, with intact relationships being the final arbiter of intervention success. The five types also differed in terms of how others involved in the intervention were characterized, the narrative form (e.g., tragedy, satire) and where and how transformative change in communities was best created. This meant that different types traded-off or managed the priorities of the intervention differently, according to the deeply held values of their type. The data set constructed for this analysis is unique. It revealed that practitioners not only exercise their agency within interventions, they do so systematically, that is, according to a pattern. The typology is the first of its kind and, if verified through replication, may have value for anticipating intervention dynamics and explaining implementation variation in community interventions.
关键词：Community Development ; Community Intervention ; Ideal Type ; Intervention Implementation ; Narrative Analysis
