Discussion

Problems addressed in action research are complex, defying simple understanding and objective interpretation \citep*{Sagor}. Some forms of action research are conducted with the purpose of clarifying a problem \citep{Putman}. These clearly do not require pre-study analysis of the problem given that the purpose of the study itself is to provide a better understanding of the problem. However, practical forms of action research, that are conducted to test the effectiveness of an intervention, do require a degree of pre-study clarity about the problem. This latter form of research is what we had in mind when developing the proposed method of problem analysis.
Problems can be articulated any number of ways depending on the researcher’s schema and the context in which the problem is situated \citep{deGrave}; therefore, a systematic and transparent process for analyzing the problem helps the researcher be clear about what is being addressed, prioritize research efforts, and communicate that articulation with others. We propose that a systematic and transparent problem analysis conducted as a pre-study task would enhance the rigor and transferability of findings from action research just as a thorough description of research methods currently does. Although we anticipate some resistance to adding another task to the already daunting process of action research, consider this: a thoroughly analyzed problem contextually framed and deconstructed into its root drivers can serve as a nexus for multiple action research projects, each addressing a specific driver. In this way, the expenditure of time with pre-study analysis may return the drivers that can each become the impetus for multiple interrelated research projects.

Analysis Method Progenitors

Several processes already exist for clarifying problems; and the Heuristic Analysis method draws on, combines, and builds upon elements from these other processes. Among these, the driver diagram and failure modes and effects analysis (FMEA) have been the most influential; and readers familiar with them will recognize their influence on the presented method. A full discussion of other methods is beyond the scope of this article, but a few words about each is warranted.
The driver diagram sees frequent use in the field of medicine, particularly medical practice and delivery  \citep{Reed,White}. Its strength is the ability to break down a complex problem into constituent levels (i.e., level 1 and level 2 drivers). This helps reveal the root causes of the problem, which encourages more targeted research. The FMEA, of which several different variations exist, sees frequent use in engineering and programming \citep{Arabian-Hoseynabadi,Ben-Daya}. Its strength is the ability to identify and prioritize points of failure within a system so the developer can address them during design and testing.
Both of these techniques can be effective for deconstructing or prioritizing a problem; but individually they fall short of fully analyzing the problem in a way that: a) recognizes the hermeneutic and situated nature of problem analysis, and b) provides sufficient clarity for contextually-situated action research. In other words, they do not individually generate a high-resolution problem: one with sufficient detail about the ambit, composition, and tenacity in a way that informs time and resource sensitive action research.
Opacity and inertia around a problem can be overcome through careful analysis that reveals its root causes \citep*{Whitehead}. Disentangling symptoms from root causes requires systematic investigation into, and analysis of, the problem so research efforts can be maximally efficient. Efficiency here means generating a practically important effect with minimal effort over the shortest period of time. To make this concrete, a high school principal working to address high drop-out rates cannot wait years for the findings from a research study before taking action to address the problem. Conversely, implementing an intervention and hoping for the best is equally untenable. What’s required is a thorough and contextually-situated understanding of the problem that clarifies its root causes.

Limitations and Conclusion

It is important to acknowledge that no clinical trials of the Heuristic Analysis process have been conducted at the time of writing. Current and earlier iterations of the process have been used in the field, but its ability to provide enhanced analysis of a problem relative to a control condition or another analysis process has still not been empirically tested. Despite their ubiquity, there is a surprising lack of efficacy or even effectiveness research for any of the common analytic methods such as those mentioned in this section. The efficacy and application of the Heuristic Analysis method and other analysis models is an area that requires further research.
In this article, we proposed a method developed to enhance problem analysis and transparency within action research. At the core of our argument is the belief that practical action research can be made more effective and leveraged for advances within the field if the analytic logic trail is clearly articulated and made transparent to other researchers and stakeholders. This requires that researchers acknowledge the subjectivity inherent in the act of analyzing a problem \citep*{Bryk}, even when that analysis is done in preparation for interventionist action research conducted from a positivist orientation. The Heuristic Analysis method was designed to fill this need and provide a living process that can be adopted outright or adapted to meet the needs of the study.