DIMACS Workshop on Evidence-based Policy Making

December 2 - 3, 2010
Lamsade, University Paris Dauphine, France

Organizers:
Fred Roberts, Rutgers University, froberts at dimacs.rutgers.edu
Alexis Tsoukiàs, LAMSADE, tsoukias at lamsade.dauphine.fr
Presented under the auspices of the Special Focus on Algorithmic Decision Theory.

Abstracts:

Olivier Barreteau, CEMAGREF

Title: Simulation model of "drought committees" in practice

The French water act institutionalizes a drought committee at county level. Such committee sets the rules characterizing a situation of drought and how to react when such situations occur. The enforcement of these rules depends on the actual characterization of a drought situation, which happens to be controversial due to the salience of the issue for participants in such committees and to the multiplicity of places and protocols to assess the actual situation. I will first describe how various information sources are considered in such committees, then the consequences on the enforcement of drought management rules supposedly based on "scientific" knowledge and evidence. In a last part I will present a simulation model aiming at exploring the sensitivity of a virtual basin to scenarios of action of such drought committees.


Cliff Behrens, Telcordia Applied Research

Title: Making Policy with Sparse Evidence: When Expert Judgments Count

There are situations when little empirical information exists for policy-making. For example, there may not be enough time or funding to gather new information, or in some extreme situations, such as anticipating the consequences of a dirty bombing in a metropolitan subway, data collection may be impossible. In cases like these, judgments provided by subject matter experts can prove invaluable, and may be the only resort. This talk will review some of the issues associated with using experts to make policies in the larger context of expert panel lifecycles. This synoptic view properly focuses attention on the selection and qualification of experts, the development of valid data-collection instruments, the application of experimental designs to proactively solicit representative data, and statistical methods for deriving consensus answers from expert panel data. The technical approach proposed in this talk takes advantage of new Web-based collaboration infrastructures and "cultural consensus analysis," a technique developed by quantitative anthropologists to derive the "best" set of answers to questions. This approach is particularly useful for stimulating idea-sharing, exposing expert biases and mitigating the adverse effects of "group think."


Marta Bottero, Politecnico di Torino

Title: Sustainability assessment and public participation in territorial planning

Sustainable development is a multidimensional concept that includes socio-economic, ecological, technical and ethical perspectives and thus leads to issues that are characterized by a high degree of conflict, complexity and uncertainty. The assessment of alternative scenarios of territorial transformation can be seen as a complex decision problem where different aspects need to be taken into account simultaneously, with reference both to technical elements, which are based on empirical observations, and non technical elements, which are based on social visions, preferences and feelings. Furthermore, when speaking about urban and territorial planning, public participation and citizens' involvement in decision-making processes have become an issue of increasing importance over the last decades. The work addresses the problem of sustainability assessment and public participation in urban and territorial planning using the Analytic Network Process (Saaty, 2005). The case study refers to the decision problem concerning the development of the Municipal Plan of Volta Mantovana (Italy). Five different zones have been identified and compared through the ANP in order to select the most important areas to which it is necessary to pay more attention in the planning process. Participation of the stakeholders is a central part of the proposed approach. As a result, the work proposes some reflections concerning the critical validation of the model and the possibilities of expanding the study.


Denis Bouyssou, CNRS - LAMSADE

Title: Ranking scientists and departments in a consistent manner

The standard data that we use when computing bibliometric rankings of scientists are just their publication / citation records, i.e., so many papers with 0 citation, so many with 1 citation, so many with 2 citations, etc. The standard data for bibliometric rankings of departments have the same structure. It is therefore tempting (and many authors gave in to temptation) to use the same method for computing rankings of scientists and rankings of departments. Depending on the method, this can yield quite surprising and unpleasant results. Suppose for instance that we have two departments each consisting of two scientists. The scientists in department 1 both have 4 papers, each one cited 4 times. The scientists in department 2 both have 3 papers, each one cited 6 times. Let us rank the scientists and the departments using the h-index. Both scientists in department 1 have h-index 4 and are therefore better than both scientists in department 2, with h-index 3. Yet, department 1 has h-index 4 and is therefore worse than department 2 with h-index 6, so that the "best" department contains the "worst" scientists. This problem will not occur if the rankings satisfy a property called consistency, recently introduced in the literature. In this paper, we explore the consequences of consistency and we characterize two families of consistent rankings.

(Joint work with Thierry Marchant).


Maria Franca Norese, Politecnico di Torino

Title: How to support and orient monitoring towards a learning system for the future

Monitoring processes can support policy making with information, knowledge and values that are not naturally available but have to be "stimulated" and then synthesised and transferred. This kind of intervention should not perturb the monitoring context and could orient the process. It has to elaborate a "picture of the situation" clear enough to be understood and collectively analysed. At the same time this picture has to be perceived as the result of a totally documented procedure and not as a personal and subjective elaboration. The integrated use of formal tools can accomplish such objectives, when combined approaches and methods support communication, visualisation of complex relations and translation of the different significant elements in a transparent, consistent and formal language. Knowledge structures emerging from case studies are proposed. They show how multiple criteria methods and problem/model structuring methodologies can be used in monitoring contexts in order to support policy makers in understanding contexts, processes and related results as well as in conceiving improvements or new policies.


Simon French, Manchester Business School

Title: Expert Judgement, Meta-Analysis and Evidence Based Policy Making

Some twenty five years ago, I distinguished three contexts in which one might wish to combine expert judgements of uncertainty: the expert problem, the group decision problem and the textbook problem. Over the intervening years much has been written on the first two, which have the focus of a single decision context, but little on the third, though the closely related field of meta-analysis has developed considerably. The text-book problem relates to how one should draw expert judgements into a decision analysis when those judgements were made originally in other contexts. However, as societal decision making has become more open, as stakeholders have become involved in framing and deciding on policy, and as the imperatives for evidence-based decision making have grown, the need to address the text-book problem has become more apparent. Further the growth of the Web and the ease with which we may all access past reports and studies have exacerbated our need for coherent methodologies to combine and use expert judgement 'out of context'. Put simply, we have meta-analyses for data; we need them for judgements too.


Thierry Marchant, Ghent University

Title: Implementable Social Choice Functions

When the domain is R and there are at least three alternatives, Roberts [1979] has proved that any implementable and neutral social choice function must be a weighted welfare maximizer. This important result has two drawbacks: (1) it does not say anything when the domain is a subset of R and (2) it does not address the case of two alternatives. Yet, these two cases are important for many applications (for instance auctions). Mishra and Sen [2010] have extended Roberts' Theorem by considering domains consisting in an open interval of R. But they need an additional condition (neutrality), yielding unweighted welfare maximizers. In this paper, we characterize the family of weighted welfare maximizers with two alternatives and any open interval of R. We do not need neutrality.

(Joint work with Debasis Mishra).


Gilberto Montibeler, London School of Economics

Title: Weak-Evidence Based Decision Making: Assessing Animal Health Threats for DEFRA

One of the pillars of evidence-based decision making is that policy makers should use the best available evidence to assess the consequences of implementing a policy. But what if the evidence required is incomplete, soft or unavailable? Even more challenging, what if there are many options to be assessed and thus extensive evidence gathering is unfeasible or too expensive? In this talk we discuss these challenges in the context of a real-world project that we recently were involved: the development of a decision support system for DEFRA (the UK Department for Environment, Food and Rural Affairs) to assess and prioritise animal health threats.


Alec Morton, London School of Economics

Title: Incorporating health inequalities considerations in healthcare priority setting.

How should commissioners and policy makers take into account health inequalities in their prioritisation of possible healthcare expenditures? In this paper, which is based on extensive experience working with commissioners in England and Canada, we present a theory-based and practical method for informing prioritisation decisions which takes explicit account of health inequalities in a way which imposes only modest time and data demands on Primary Care Trusts and their management.


Gabriella Pigozzi, LAMSADE

Title: Judgment aggregation in abstract argumentation

Individuals may draw different conclusions from the same evidence. For example, members of a jury may disagree on the verdict even though each member possesses the same evidence regarding the case under discussion. This happens because individuals can hold different reasonable positions based on the information they share.

The field of judgment aggregation studies how individual positions on the same information can be aggregated into a collective one. After a gentle introduction to judgment aggregation, I will offer an analysis of judgment aggregation problems using an argumentation approach. One of the principles of argumentation theory is that an argumentation framework can have several labellings. If the information the group shares is represented by an argumentation framework, and each agent's reasonable position is a labelling of that argumentation framework, the question becomes how to aggregate the individual positions into a collective one.

Whereas judgment aggregation focuses on the observation that the aggregation of individual logically consistent judgments may lead to an inconsistent group outcome, I will present an approach that not only ensures collective rationality but also social outcomes that are 'compatible' with the individuals' evaluations. This ensures that no individual member has to become committed to a group position that is in conflict with his own individual position.

(Part of my presentation will be based on a joint work with Martin Caminada and Mikolaj Podlaszewski.)


Marc Pirlot, Université de Mons

Title: Decision aiding in geographical contexts

Conceiving and developing tools for aiding decision in problems amenable to spatial representation has received increasing attention in recent years. Applications pertain, for instance, to land use management and localization of facilities (windfarms, obnoxious facilities,...). One of the goals is to implement decision aiding tools and methods in geographic information systems (GIS). In this presentation, we focus on new methods developed in the context of an application to land use assessment of a region in West Africa (Burkina Faso) in the perspective of sustainable development. We mainly address the question of comparing the state of a territory at two different moments. This question is of importance for the assessment of policies that aim at improving the current state of a given region. We propose two models allowing for such an assessment on the basis of observing the current state of a region and the way it has evolved after a period of application of a policy.


David Rios Insua, Royal Academy of Sciences

Title: Evidence based policy making: Some experiences from water management problems

Water resources problems provide many challenges for policy making due to the presence of multiple objectives, multiple uncertainties, several decision makers, multiple stakeholders and the effects of time. We shall describe two complex water resources problems in which we have been recently involved:

We shall cover some of the lessons learnt in dealing with such problems.

(work done with Angel Udias, Javier Cano, Universidad Rey Juan Carlos, Kiombo Jean Marie, University Agostinho Neto, Hocine Fellag, Université Tizi Ouzou)


Fred Roberts, DIMACS

Title: Measurement of Biodiversity

Among the potential effects of climate change, health of ecosystems is of concern. Evidence for the health of ecosystems is often obtained by measuring the biodiversity. An index of biodiversity allows us to set specific goals and measure progress toward them. This talk will survey indices for biodiversity and axiomatic approaches leading to such indices.


Alexis Tsoukiàs, CNRS - LAMSADE,

Title: What is evident in evidence based policy making?

The necessity to use some evidence in order to conceive, design and/or assess public policies became a subject of debate already several years ago. At the same time it started to become a subject of scientific investigation. However, it is not really clear what evidence means in supporting a policy. First of all what we consider as "facts" in reality are synthetic measurements of more complex realities and as such subject to biases and uncertainties of any measurement procedure. Besides, quite often such information has to be extracted out of huge data bases and or streaming processes. Second, such facts need to be understood and interpreted in a shared way by different stakeholders and it is not clear how such a process can occur consistently. Third, policies are also based on values, preferences and judgments (possibly representing large parties of the societies affected by the policies) and the way to consider these can again be a subject of controversy. Fourth, the way through which all such information is manipulated can end up constructing "different evidences" by just adopting (for instance) a different concept of majority. Based on such remarks we elaborate the claim that evidence based policy making should be understood as a specific type of decision aiding process. We then introduce a second claim: the challenge for decision sciences and technologies in this area consists in elaborating a "policy analytics" perspective.

(Part of my presentation uses work done with Wassila Ouerdane, Giulia Lucertini and Giada de Marchi).


Decision aiding in geographical contexts Abstract: Conceiving and developing tools for aiding decision in problems amenable to spatial representation has received increasing attention in recent years. Applications pertain, for instance, to land use management and localization of facilities (windfarms, obnoxious facilities,...). One of the goals is to implement decision aiding tools and methods in geographic information systems (GIS). In this presentation, we focus on new methods developed in the context of an application to land use assessment of a region in West Africa (Burkina Faso) in the perspective of sustainable development. We mainly address the question of comparing the state of a territory at two different moments. This question is of importance for the assessment of policies that aim at improving the current state of a given region. We propose two models allowing for such an assessment on the basis of observing the current state of a region and the way it has evolved after a period of application of a policy.
Previous: Program
Workshop Index
DIMACS Homepage
Contacting the Center
Document last modified on December 8, 2008.