London Judgement and Decision Making seminars
The LJDM seminar series is supported by
University College London
City University London
Originally established at UCL in the early 1970’s as a weekly Cognition and Reasoning seminar, it later became an intercollegiate seminar on Language and Cognition in the early 1980’s.
The name LJDM was finally coined in 1990, and the group has been running seminars under this name ever since, with lecturers and researchers in and around the UK meeting on a regular basis to discuss judgment and decision making, judgments of likelihood, reasoning, thinking, problem solving, forecasting, risk perception and communication, and other related topics.
If you would like to present your research to the group or to suggest a speaker, please contact the organizers, Neil Bramley (email@example.com), Leonardo Cohen (firstname.lastname@example.org) or Eric Schulz (email@example.com).
Unless specified otherwise, all seminars take place on Wednesdays at 5pm, in Room 313 at the Psychology Department, University College London (on the corner of Bedford Way, Gordon Square and Torrington Place, London WC1H 0AP). Map.
To get updates on the current schedule and weekly reminders of the seminars, please subscribe to the Risk and Decision mailing list.
All are welcome to attend.
Current Seminar Schedule
October – December 2015
What drives “Unconscious” Multi-Attribute Decision-Making?
Birkbeck, University of London
This study aims to further investigate the Unconscious Thought Theory (UTT, Dijksterhuis & Nordgren, 2006), namely whether individual differences account for differences in choice made after either deliberation (conscious thought, CT) or distraction (“unconscious thought”, UT). Subjective weighting was taken into consideration by providing choice options that were constructed following individual preferences. Hence attributes match the subjective preferences, which undermines previous critiques that results were distorted by differences between individual and objective choice. The main effect was replicated with a big sample (N=120, CT: 50.8%, UT: 70.5% chose the best alternative), using four different dependent measures. This result is remarkable insofar as Nieuwenstein and van Rijn (2012) in their meta-analysis only found an effect when CT was at chance level. The results show further that the main effect is driven by underperformance of women in the CT condition. Stereotype threat is discussed as a possible explanation.
Time compounding not time discounting
University of Stirling
Intertemporal choice involves comparing prospects in different time periods. Instead of discounting future values to present values, intertemporal choice can proceed by compounding present values to future values. Maximizing compounded future value is equivalent to maximizing discounted present value. Time compounding is presented as a more plausible model of processes underlying intertemporal choice than time discounting models. Time compounding leads to simple graphical analogues for impulsive behaviours. Linear extrapolation of growth in future value is proposed to explain within-person variation in impulsiveness. Impulsive behaviour is driven by linear overestimation of logarithmic value, and linear underestimation of exponential value. Time compounding incorporates additional empirical regularities in decision-making.
Presenting Bayesian problems to a general audience
Queen Mary, University of London
The ability to undertake Bayesian inference is becoming a necessary skill in modern society, particularly within medicine and law. In medicine, doctors and patients need Bayesian inference to make accurate assessments of the risk of a patient having a condition given several pieces of evidence. In law, jurors are increasingly being presented with Bayesian calculations with the expectation that they will be able to comprehend the information to provide a fair and accurate judgement. However, both professionals and the public have been shown to make large errors in Bayesian estimations. Two approaches have had success in improving accuracy on Bayesian problems such as those faced by these groups: the ‘Nested Sets’ and ‘Causal’ framing approaches. In a first paper, these approaches were combined and compared on a medical problem. Further, participants were encouraged to document their thought processes while solving the problem using a ‘write aloud’ protocol. An improvement in accuracy was seen for the nested sets framing, but no effect was seen for the causal framing. From the ‘write aloud’ data a 4-stage model was developed demonstrating the process that all successful individuals undertake. In a second paper this nested sets framing approach was tested on a range of more ecologically valid problems and was still found to improve accuracy. There are concerns that this approach will not be feasible in legal situations due to its inherent focus on a single individual (the defendant), and so in a prospective paper, this ‘problem of law’ will be tackled.
Staying afloat on Neurath’s boat: A new theory of theory change
University College London
Over their lifetimes, people develop rich causal representations of the world that allow them to understand and flexibly exploit their environment. People have been shown to be effective causal learners, yet rational accounts of causal structure learning are intractable for all but the smallest of toy problems. To account for this discrepancy, I propose a novel scalable process model of active causal learning. The model is inspired by an old idea from philosophy of science about the sequential and piecemeal nature of theory change, but also draws on current ideas from machine learning, about sequential and approximate techniques for approximating Bayesian updating. I will show that our model better captures the ways that people gather evidence and update their causal beliefs than a range of alternatives in two causal learning experiments. Finally, I will suggest that the model is not limited to the domain of causal structure learning, but is a candidate explanation for the gradual evolution and adaptation of many types of theories in cognition.
Justifying unpopular climate policies: The argument from moral error and moral expertise
University of Warwick
If humankind is to avoid disastrous climate change, rapid and drastic changes to many aspects of our economy and lifestyles are needed. Some policies that are needed to implement such changes are going to be very unpopular, most notably if they incur non-monetary cost such as lifestyle and land-usage alterations, and if such cost cannot be distributed equitably. This situation raises the normative question of whether policy-makers are justified in even trying to implement such policies, e.g. by using obfuscation, financial lock-in, and other means of insulating these policies against democratic opposition. In this paper, I provide a justification of such policies on grounds that opposition to these policies arises from a specific form of moral error that can justifiably be compensated by moral experts.
In particular, I argue that the moral decision situation by posed climate change has several features which undermine the quality of our everyday moral decision-making. I further contend that moral philosopher's specific expertise is to understand and compensate for these distorting features. Policies which are based on the advice of such moral experts then do not disrespect, and, if coupled with education and social change, even enhance the autonomy of citizens. Because my proposal rests on a highly specific understanding of moral expertise, it avoids justifying an implausibly wide and intrusive range of ethics-driven policies.
Designing for the psychology underpinning behaviour change
Queen Mary, University of London
Knowledge of the psychological constructs that underlie our actions offers valuable design opportunities for behaviour change systems. I discuss three projects that aim to uncover these constructs. First I will discuss how we applied decision theory based analysis in the domain of weight management to understand the rewards and costs that surround individuals’ weight management behaviours. Next, I will discuss a study that uses behavioural economics methods to evaluate a major psychological cost of weight management: the perceived cost of effort for sitting, standing, and walking. Finally, I will discuss a mobile app we designed to help people with the psychological challenge of desire for excessive and unhealthy food. Our mobile app delivers a food craving reduction intervention at the moment of need and allows users to track how often they successfully resisted cravings. The craving reduction intervention used is based on recent research that has found that food cravings can be reduced through imagery techniques. We conducted a week-long evaluation of our app, comparing its effectiveness to a basic tracking application. We found that our imagery application significantly reduced both overall snacking and unhealthy snacking compared to a simple snack-tracking application.
A Model of Factors that Shape Cancer Screening Decisions
University of Granada/Imperial College London
Contrary to people’s intuitions, many cancer screenings have both benefits and harms. For example, screening can reduce cancer mortality but result in many unnecessary surgeries. Statistical information about such benefits and harms is sometimes provided to patients to foster informed decision making. However, few models have addressed the role of comprehension of such information in screening decisions. So how do people decide about screening when it is not only good? How do they weigh in the evidence deemed relevant by experts? In three experiments, we studied how several cognitive and emotional factors affect patients’ decisions. Our results suggest that risk literacy (e.g., numeracy, science literacy) and a healthy dose of fear can promote informed decision making. At the same time, emotions and beliefs reinforced by persuasive campaigns can have strong effects on patients’ decisions beyond the available evidence. To provide a more thorough picture of the decision environment, in a fourth experiment, we investigate how physicians communicate about screening with their patients and what factors shape their recommendations. Potential mechanisms (e.g., motivated cognition) and implications for informed decision making will be discussed.
Situating a Bayesian source credibility model: initial explorations of public health campaigns and election candidates
Birkbeck, University of London
Bayesian models of source credibility have been explored conceptually (e.g. Bovens & Hartmann, 2003; Hahn et al., 2009) and tested empirically (Harris et al., in press), defining source credibility as trustworthiness and epistemic expertise. The talk discusses two concurrent aspects of the above model. Firstly, it applies the model to a real-life situation. It tests the model against appeals to specific presidential candidates. Eliciting estimations for all relevant aspects of the model, participants read a dialogue involving one of five candidates (Sanders, Clinton, Rubio, Bush, and Trump) and rated how strongly they believed a proposed policy to be good if that candidate publically had attacked/supported the policy. Secondly, factorial analyses indicate that, like expertise, trustworthiness may be domain and context-dependent (Madsen, submitted). Taking point of departure in analyses of doctors from public health campaigns in the UK and Denmark, the concept of trust in the original model is expanded to a multi-facetted concept (‘Ethical behaviour’, ‘Care for the health of all people’, and ‘altruistic diligence’ in the UK). This provides a test of a highly detailed and context-dependent source description. The talk provides indicative evidence that the general model tested in Harris et al. (in press) can be applied successfully to more realistic argumentation scenarios (e.g. believing the goodness of a policy if a trusted and expert election candidate advocates it) and more complex source depictions. Although in its infancy and thus only indicative and tentative, the results point to interesting and critical applications of the Bayesian source credibility model.
Game show economics
Martijn van den Assem
Vrije Universiteit Amsterdam
Much of what we know about individual decision making to date derives from laboratory experiments, and less so from real-life settings. Real-world data typically entail a lack of control, which can make it difficult to discriminate between competing hypotheses. Carefully designed laboratory experiments do allow for this and have generated a rich literature, but this approach is also vulnerable to criticism about the generalizability of findings to situations beyond the context of the lab. Because it is impossible to study behaviour under each and every possible set of conditions, the optimal approach to assess the generalizability of existing findings is to study behaviour in a limited number of diverging ways. The use of TV game shows is one of these. Game shows allow the study of behaviour in a high-scrutiny field setting where the stakes are high, and for a diverse subject pool. Combined with the strict and well-defined rules, game shows can provide unique opportunities to investigate the robustness of existing laboratory findings. In my talk, I will present my game show work on bargaining, risk taking, and cooperation.
The flexibility of moral intuitions
Jonas Nagel and Alexander Wiegmann
University of Goettingen
Moral intuitions play an important role in philosophical and everyday moral discourse. Often they serve as starting point or ultimate justification in moral arguments. In several different lines of research, we demonstrate the pronounced context-sensitivity and flexibility of people’s moral intuitions in response to concrete case descriptions. We conclude that the expressed judgments are not well characterized as the accurate expression of stable underlying moral values, but rather reflect the output of a highly flexible process of largely domain-general reason-based judgment and choice.