Hierarchy of Hypotheses as an organizing tool for research. Report on a workshop

Journal of Ecology Executive Editor David Gibson, Associate Editor Lorena Gómez-Aparicio, and Methods in Ecology and Evolution Senior Editor Bob O’Hara recently attended a workshop “The hierarchy-of-hypotheses approach: Exploring its potential for structuring and analyzing theory, research and evidence across disciplines” August 19-21, in Hannover, Germany. Below they tell us more about this workshop.

---

Set in the splendid Schloss Herrenhausen in Hannover, Germany, the workshop was attended by 28 participants that included ecologists, and experts from other disciplines including philosophers of science, oncology and chemistry. Sponsored by VolkswagenStiftung, the goal of the workshop was to discuss and investigate the potential for broad application of the Hierarchy of Hypotheses (HoH) approach.

Schloss Herrenhausen setting for the Workshop, and workshop participants.

Developed by workshop organizers Tina Heger and Jonathan Jeschke to compare competing hypotheses in invasion ecology, the HoH approach highlights the need of hierarchically structuring major (usually very general) hypotheses into smaller more specific sub-hypotheses that can be empirically tested (see figure below).

The workshop was organized around a number of themes, namely explanations of HoH, alternative approaches (e.g., Self-organizing Hypothesis Networks from chemistry), potential applications in other disciplines beyond invasion biology including paleoecology, population biology (see David Gibson’s presentation), oncology and environmental policy, possible ways to improve the HoH framework and analysis using approaches such as the use of meta-analyses, expert analysis, and evidence assessment tools.

Example of a Hierarchy of Hypotheses for invasion ecology. Overarching ideas branch into more precise, better testable hypotheses at lower levels. Empty boxes indicate that the hierarchy may be extended. Reprinted in Gibson (2015) from Heger et al., (2013).

David Gibson writes, “My view is that the workshop was very focused, which was a welcome change from the smorgasbord of concurrent sessions at the large national meetings that I usually attend. Things got lively during discussions and in breakout groups where we discussed the various pros- and cons of topics arising from the presentations especially in relation to areas where HoH was viewed as imprecise, unclear, or subjective. Issues that generated much discussion included the different ways to apply the HoH approach, how to construct the hierarchy for a set of hypotheses in the first place, the choice of branching criteria, whether all levels in the hierarchy should even be hypotheses (or research questions, or goals), when and how to weigh individual studies when comparing hypotheses (something that Jeschke emphasized was optional anyway), the subjectivity involved in supporting or rejecting hypotheses, and possibilities for combining HoH with other synthesis tools. Even on the last morning of the workshop there was still disagreement among some participants, but the airing of views will be useful for the organizers moving forward. Personally, I became more enthusiastic about the HoH over the course of the two and half days, and feel that the semantic and conceptual issues raised can be adequately clarified. The group agreed that the next steps are to write manuscripts to promote HoH to other disciplines beyond invasion ecology, and to demonstrate the utility of incorporating additional synthesis tools.”

Lorena Gómez-Aparicio writes, “It was really stimulating to have the opportunity to share opinions with people with very different backgrounds about the multiple ways scientists can look for generality and knowledge synthesis. Although I’m a big fan of meta-analysis myself, workshops like this remind me of the multiple tools available out there to organize empirical evidence in a useful way for advancing our ability to explain and predict ecological processes. I find the combination of several of these techniques, like for example HoH and meta-analysis, the most promising avenue for research synthesis. I also enjoyed very much the possibility of discovering totally new initiatives to organize scientific knowledge such as the visual interface Open Knowledge Maps (see Peter Kraker’s presentation). This workshop was intended to be the first one of a series of three. I hope that the high satisfaction of all the attendants after our three days in Hannover serves as a proof of concept that organizing those two additional workshops in the future is really worthy”.

Bob O’Hara writes, “The HoH approach tries to do two things: first, it tries to organize theory, by building a hierarchy of theories/hypotheses/mechanisms (or whatever – that was one discussion during the workshop). Second, it tries to synthesize the evidence for the different parts of the hierarchy of hypotheses. My own feeling was that the first aspect was well received, but the evidence for synthesis had some problems. But the second day of the workshop was focused on evidence synthesis, and different ways of combining these methods with HoH would seem to be a fruitful area to be explored. What I really like about this approach is that it provides a framework to systematise theories, and make the link between high-level ideas (e.g. the Red Queen Hypothesis), the way this general idea is made.”

David Gibson, Executive Editor