Post provided by Blal Adem Esmail & Davide Geneletti

Comparing Apples and Oranges

©Ruth Hartnup

In real-life situations, it is far more common for decisions to be based on a comparison between things that can’t be judged on the same standards. Whether you’re choosing a dish or a house or an area to prioritise for conservation you need to weigh up completely different things like cost, size, feasibility, acceptability, and desirability.

Those three examples of decisions differ in terms of complexity – you’d need specific expert knowledge and/or the involvement of other key stakeholders to choose conservation prioritisation areas, but probably not to pick a dish. The bottom line is they all require evaluating different alternatives to achieve the desired goal. This is the essence of multi-criteria decision analysis (MCDA). In MCDA the pros and cons of different alternatives are assessed against a number of diverse, yet clearly defined, criteria. Interestingly, the criteria can be expressed in different units, including monetary, biophysical, or simply qualitative terms.

Three Main Stages of Multicriteria Decision Analysis

In general, MCDA can be divided into three main stages, as described in our recent paper in Methods in Ecology and Evolution: ‘Multi‐criteria decision analysis for nature conservation: A review of 20 years of applications’. In the first stage, you need to establish a shared understanding of the decision context and structure the problem by involving key stakeholders and experts. The best way to do this is by:

  • Clearly defining the objectives of the decision process
  • Identifying possible alternatives to achieve them
  • And formulating explicit criteria to assess how each one contributes to the objectives

The second stage is the actual analysis. It includes criteria assessment (i.e. evaluating how well each alternative performs based a dimensionless scale of desirability defined by experts and/or stakeholders), relative weighting of the different criteria (i.e. willingness to trade-off between the criteria), and then an aggregation of the performances across the criteria. A key step here is the sensitivity analysis to test the robustness of the results. Finally, in the third stage of the MCDA you bring together information from the previous stages to rank the alternatives, ultimately leading to the actual decision.

Our Review, its Main Findings, and Conclusions

Focusing on applications in conservation science, we wanted to investigate how the literature has addressed the three main stages of the MCDA process. We were particularly interested in how people had described the level of involvement of different actors and the variety of techniques that have been applied (such as Delphi technique, interviews, focus group discussions, and nominal group techniques).

Thorough a structured keyword search, we identified a representative sample of 86 articles, classified by their own authors as explicitly addressing conservation issues. We analysed these articles in the light of best practices and common pitfalls reported in the MCDA literature with the aim of providing recommendations to future MCDA applications in nature conservation.

Most of the reviewed articles addressed forest management and restoration (25%), conservation prioritization and planning (24%), and protected area planning and management (21%), while fewer dealt with mapping of biodiversity, wilderness and naturalness (8%). The majority of the studies were carried out at either a regional level (25%) or local level (33%).

25% of articles that we analysed used MCDA to study forest management. ©Chiara Cortinovis
25% of articles that we analysed used MCDA to study forest management. ©Chiara Cortinovis

A few common pitfalls emerged from our review. One of the main ones was poor structuring. In general, this was because of unrepresentative sets of alternatives, excessive and unbalanced criteria for different objectives, and (in some cases) minimal involvement of stakeholders. Also, most of the reviewed articles didn’t explicitly discuss the rationale behind the criteria assessment step in the analysis step. They didn’t specify how the dimensionless scale of desirability of each criterion was obtained. This is a common pitfall of MCDA applications – little attention is given to how information about performance of each alternative is converted into a dimensionless scale of preference that expresses the level of desirability of that alternative.

Adequate understanding about the implication of weights is another crucial aspect of a successful MCDA; but, in the reviewed articles, such information was often scarcely conveyed to the key stakeholders and experts. Weights should be assigned to account for the different degrees of importance of the criteria to the decision, considering different stakeholders’ perspective. Finally, a successful MCDA application should always include a sensitivity analysis to examine the trustworthiness and, unfortunately, not all of the sampled articles included one.

Our review showed that MCDA applications have not always been used to their full potential in conservation research. MCDA is a suitable and flexible approach though and, if used properly, could be really useful in this field. We hope that our recommendations encourage more people to use this approach to improve decision-making in nature conservation.

To find out more about Multi-Criteria Decision Analysis in conservation research, read our Methods in Ecology and Evolution article Multi‐criteria decision analysis for nature conservation: A review of 20 years of applications’.

This article is part of the ‘Qualitative methods for eliciting judgements for decision making’ Special Feature. All articles in this Special Feature will be free throughout 2018.