In attempting to use a realistic evaluation approach to explore the role of Community Parents in early parenting programs in Toronto, a novel technique was developed to analyze the links between contexts (C), mechanisms (M) and outcomes (O) directly from experienced practitioner interviews. Rather than coding the interviews into themes in terms of context, intervention elements (mechanisms) and outcomes separately and which could be assembled into CMO configurations by the analyst, they were coded as linked dyads and triads directly from the practitioner narratives. Out of all of the linked codes entered, there were a maximum of three with the same combination, presenting challenges for typical qualitative data analysis. This article examines a novel technique that was developed in an attempt to expand this method beyond the circumstances described in the realistic evaluation literature to date. The bulk of the article focuses on the linked coding and analysis procedures, the challenges faced, and the original solutions that were developed to analyze the CMO relations and generate the mid-range theories necessary to move to the next stage of a realist evaluation approach. The features that distinguish this linked coding method from other methods (e.g. Qualitative Comparative Analysis), the major benefits and drawbacks, the utility of the approach within evaluation practice, and its application to realist synthesis and research are discussed. Downloaded from and published evidence, the literature typically describes a qualitative interview process with the plausible hypotheses derived logically from the data. Pawson and Tilley describe an interactive interview or dialogue as a method to understand context, mechanism, and outcome relations where the interviewer and the interviewee co-develop these hypotheses (Pawson & Tilley, 1997). If many practitioners are involved, is there another way that these relations can be uncovered using interviews with a group of practitioners?Several methodological problems are posed in trying to do this. One is the need to ask the right questions to collect information about the contexts, and the program activities and characteristics that could become mechanisms. A second is how to recognize the CMO connections that are made implicitly or explicitly in the data. A third major problem is how to analyze connected strings of data in a way that is different from grouping codes under common themes. This article will examine a novel technique that was developed in an attempt to expand this method beyond the circumstances described in the realistic evaluation literature to date. We begin by examining the realist evaluation literature related to coding, and then describe the case and data collection methods. The bulk of the article focuses on the analysis methods, the challenges we faced, and the solutions that were developed. We end by describing the strengths and limitations of the linked coding approach and suggest implications for evaluation practice and future research.