Abteilung Umweltsozialwissenschaften

Complex environmental and public policy decisions profit from structured procedures such as multi-criteria decision analysis (MCDA). To support such decisions, the new open source application ValueDecisions provides advanced analysis and visualization with no programming expected from users. Based on multi-attribute value theory (MAVT), it offers analysis for decisions with conflicting and interacting objectives, multiple stakeholders, and uncertain consequences of options. Programmed in R, the shiny web framework makes it accessible via a graphical user interface in the browser. We exemplify using ValueDecisions for a wastewater infrastructure planning case in the Paris region. We surveyed preferences of 655 citizens and conducted sensitivity analysis of preference parameters. The best management options were robust across a range of preference profiles and assumptions. To evaluate the app, we developed a novel usability test based on the ISO standard for software quality and surveyed students using ValueDecisions for case studies.

Decision-making in environmental management requires eliciting preferences of stakeholders and predicting outcomes of decision alternatives. Usually, preferences and predictions are both uncertain. Uncertainty of predictions can be tackled by multi-attribute utility theory, but the uncertainty of preferences remains a challenge. We demonstrate an approach for including both uncertainties in a multi-criteria decision analysis (MCDA), using utility theory and the concept of expected expected utility. For a decision regarding a regional merger of wastewater infrastructure in Switzerland, we constructed preference models for four stakeholders. These models also allowed for non-additive interactions between objectives. We evaluated the performance of eleven decision alternatives for which we predicted potential outcomes. Even though uncertainties were high, we could draw conclusions based on the expected expected utility of alternatives. Building a pipeline to discharge treated wastewater to a larger river emerged as a potential consensus alternative to mitigate the problem of micropollutants in a small stream. We investigated the robustness of the findings with sensitivity analysis regarding the preference parameters and the included objectives. In their actual decision, the stakeholders partly preferred other alternatives than those proposed by the model. Their choices could be explained by reduced decision models in which only few objectives were included. This may indicate the use of simplified choice heuristics by the stakeholders. The presented approach is feasible for supporting other difficult environmental or engineering decisions in practice, for which we give a number of recommendations.

Identifying objectives is essential for decision making, but individuals have difficulties stating their important objectives. In public and environmental decisions, the diverse views of stakeholders should be included, but eliciting a broad set of objectives is challenging. We (1) study the effectiveness of individual brainstorming for eliciting objectives in a real-world setting; (2) test three interventions to support individuals in generating objectives; (3) investigate which and how many stakeholders are necessary to generate a comprehensive set of objectives; and (4) develop a feasible elicitation procedure for practice. In an experimental test, 71 stakeholders participated in five decisions about regional wastewater infrastructure planning in Switzerland. Three interventions were tested with an online survey procedure: (i) providing category cues, (ii) a perspective-taking task, and (iii) providing a predefined master list of objectives. In simple brainstorming, participants stated few objectives (M = 3.3) associated with 2.9 different categories on average. Participants consistently missed objectives they later considered important. Providing a master list was the only intervention that substantially increased the number and breadth of objectives (M = 12 objectives in M = 5.3 categories). With the help of our survey, participants generated between 30 and 38 distinct objectives for each decision case. Between five and nine participants were sufficient to generate these; more participants did not contribute new objectives. Most decision makers need help generating their objectives; combining simple brainstorming with a master list is a straightforward improvement that does not require a facilitator. An online process is promising for involving a large group of stakeholders.

Multi-criteria decision analysis (MCDA) requires an accurate representation of the preferences of decision-makers, for instance in the form of a multi-attribute value function. Typically, additivity or other stringent assumptions about the preferences are made to facilitate elicitation by assuming a simple parametric form. When relaxing such assumptions, parameters cannot be elicited easily with standard methods. We present a novel approach for identifying multi-attribute value functions which can have any shape. As preference information indifference statements are used that can be elicited by trade-off questions. Instead of asking one indifference statement for each pair of attributes, we ask for multiple trade-offs at different points in the attribute space. This allows inferring parameters of complex value functions despite the simplicity of the preference statements. Parameters are estimated by taking into account preference and elicitation uncertainty with a probabilistic model. Statistical inference supports identifying the most adequate preference model out of several candidate models through quantifying the uncertainty and assessing the need for non-additivity. The approach is elaborated for determining value functions by hierarchical aggregation. We apply it to an assessment of the ecological state of rivers, which is used to support environmental management decisions in Switzerland. Preference models of four experts were quantified, confirming the feasibility of the approach and the relevance of considering non-additive functions. The method suggests a promising direction for improving the representation of preferences.

Building a well-structured objectives hierarchy is central to multi-criteria decision analysis (MCDA). However, in the absence of a systematic methodology to support the process, this task has been described as "more art than science". Objectives hierarchies often tend to become large and constraining the size of a hierarchy can be challenging. This paper proposes and illustrates the use of a set of methods to support the simplification of the hierarchies in contexts that are "data rich" and characterised by many objectives. The aim of using the proposed approach is to support decision analysts in developing an appropriately concise decision model for the further interactions with the stakeholders. Using data from two completed environmental cases we show retrospectively how qualitative (means-ends networks), semi-quantitative (relevancy analysis) and quantitative (correlation analysis, principal component analysis, local sensitivity analysis of weights) methods, used alone or in combination, can inform hierarchy development. We evaluate the potential benefits and challenges of each method and discuss the advantages and disadvantages of the simplification of an objectives hierarchy. Questionnaire-based relevancy analysis can be a useful method to identify and communicate important objectives in the early phases of an MCDA process with stakeholders, while correlation analysis can help to identify overlapping objectives, particularly in cases having many objectives and alternatives. It is intended that the methods support a facilitator in developing a clear understanding of the problem and also prompt deeper thinking about and discussion of the appropriate structure and content of an objectives hierarchy with the stakeholders involved.