ARCHIVED - Section 1 Case Study Methodology to Understand ActNowBC as a Natural Experiment

This is a qualitative case study. Case studies conduct a holistic analysis of a social phenomenon through understanding actions, events, contexts, and processes from multiple perspectives. [3] [5] This case study on ActNowBC is both descriptive and explanatory. We describe the history, structures, and processes behind ActNowBC and analyse in depth the factors critical to its success evident thus far.

We view the ActNowBC initiative as a natural experiment. The BC government has developed an evaluation strategy for ActNowBC based on a before-and-after design with interim progress reports. This classic approach to evaluation, while essential, is often marred by concerns about the challenges of measuring outcomes and determining attribution. In measuring outcomes for ActNowBC, the challenge is to detect change in the prevalence of chronic disease risk factors at the provincial level over a relatively short period of time, that is, between 2003 and 2010. Even if the prevalence of risk factors changes, the challenge remains to attribute these outcomes to the ActNowBC initiative. New approaches in program evaluation, like outcome mapping and contribution analysis, have emphasized the necessity of building a “credible performance story” over time. [6] The first step is to demonstrate change in people and organizations targeted by the proposed intervention(s). In the case of ActNowBC, this translates into determining whether ActNowBC changes the way governmental agencies and NGOs, and people within these organizations, “do business.” This case study contributes to building the performance story for ActNowBC.

Data Collection

We obtained data for our report from two sources: 1) literature on intersectoral action for health and whole-of-government approaches; and 2) in-depth interviews with public servants, politicians, academics, and NGO officials.

Information was obtained from respondents using a semi-structured interview guide. While most interviews were conducted face-to-face, some were conducted by telephone. Interviews lasted on average approximately 60 minutes. All respondents signed informed consent forms before the interviews, indicating they understood why their participation in the study was requested. The Health Canada Ethics Board approved the study protocol in October 2007.

Selection of Respondents

We used a purposeful sampling strategy (“snowball sampling”) to identify respondents with the most information, expertise, and technical knowledge relevant to the case study. [7] We wanted to obtain a comprehensive description of ActNowBC and a diversity of opinions about its implementation, added value, and other features. The first round of interviews involved key leaders of the ActNowBC initiative, who subsequently suggested the names of other individuals we should approach. Those individuals who were recommended were interviewed, and they in turn identified other potential respondents. We conducted a total of 49 interviews with key respondents working in government (37), in academia (2), and in the NGO sector (10).

Data Analysis

All interviews were audio-recorded and transcribed. The transcripts were coded and analysed with NVivo software, version 7. [8] The coding tree was developed in steps. The analysis involved first using an open coding strategy to identify emerging themes and categories. [9] Themes and categories generated after performing line-by-line analysis of the first few transcripts were tested and further explored in subsequent interviews until saturation was reached. We used the immersion/crystallization approach [10] as the overall analytical strategy to identify themes and patterns emerging from the empirical data set.