Mental Models Interviews and Comparative Analysis
The last element of research for this case study includes the mental model interviews and subsequent comparative analysis, which is driven and influenced by the literature review and expert models discussed earlier. The interview protocol was constructed using influences from the expert model in order to fashion a semistructured interview protocol. The goal of these interviews was to gain a better understanding of the specific drivers on the quality of USACE flood risk management alongside an improved perspective on the differences in flood risk management belief and opinion between USACE planners and engineers, respectively. Additionally, we were particularly interested in determining the differences between these two groups regarding the drivers they emphasize or place particular value upon.
To accomplish these goals, we interviewed USACE engineers (n = 10), planners (n = 8), and program managers (n = 4). These interviews were conducted over the telephone. Due to the similarities in mission and job function, we considered the program managers in this case to fall under the “planner” category. Overall, the interview subjects represented a convenience sample of senior USACE personnel holding positions in various units of the Corps.
The interview process and protocol were generated by using previous work on USACE personnel beliefs and opinions in the flood risk management process as an example. Changes to this older protocol were minor. Due to the general consistency of earlier engineer participants from an October 2008 interview survey, data from five engineers is included in our analysis of this case study. We were particularly interested in information related to :
- 1. The most important drivers of US ACE flood risk management.
- 2. The scope and quality of US ACE flood risk management activities.
- 3. Drivers of the quality of collaboration, coordination, and communication.
- (a) Internally within USACE.
- (b) Externally with other government, NGO, and lay stakeholders.
With these lines of inquiry in mind, interviewers asked questions to assess key features, opinions, and perceptions of the factors listed earlier by USACE personnel as related to flood risk management (Yoe 1993). To capture a full response, questions were worded broadly for an individual topic initially, and then narrowed and focused on a series of specific elements associated with the topic. This interview style and structure is consistent with past interview efforts developed by expert model diagrams (Morgan et al. 2002; Thorne et al. 2005). Interviews lasted approximately 1 h on average (mean 66 min; range 45-82 min) and were collected from October 2008 through June 2010. All discussions were recorded for future reference and subsequently transcribed.
These interviews were then analyzed via content analysis, whereby questions and their responses were reviewed individually and assessed for common themes across respondents. For consistency, each response was put through a two-pass process. In the first pass, each response by each individual respondent was read and associated with particular elements of the Expert Model. This facilitated comparison of expert model topics across all interviews and all interview questions. The second pass included a more thorough analysis, where each response was analyzed for more fundamental themes that could be identified within topic areas associated with the expert model. We were particularly interested in relative frequencies across participants that discussed a particular theme. Due to the sample size, differences between cohorts are reported here only if they are greater than 20 % (coding is rounded to the nearest 5 % by topic frequency).