CDR at Subjective Probability, Utility and Decision Making (SPUDM) 2019
Wandi Bruine de Bruin, Gabriella Eriksson, Astrid Kause, Manos Konstantinidis, John Maule, Simon McNair, Yasmina Okan and Barbara Summers were all involved in presentations and posters at Subjective Probability, Utility and Decision Making (SPUDM) in Amsterdam on 18-22 August. Abstracts for the work are shown following.
Bruine de Bruin, Wandi, Co-authors: Simon McNair and Jonell Strough
“Age differences in anticipated emotions for future life events”
E-mail: w.bruinedebruin@leeds.ac.uk
People of all ages face decisions about their future, which involves anticipating emotional responses to events that are yet to occur. Studies with college students have shown relatively strong anticipated emotional responses to future life events, but adult age differences have been unexplored. We recruited a national UK life span sample (N=2089) to examine age differences in anticipated emotions for future life events of negative, positive, or undefined valence, and in strategies for regulating those emotions. Older adults anticipated less negative emotions than younger adults independent of event valence, but we found no age differences in anticipated positive emotions. Reports of anticipated negative and positive emotions suggested that participants expected to feel better after positive (vs. undefined) and worse after negative (vs. undefined) events. Older adults reported engaging more in functional emotion regulation strategies (e.g., savoring), and less in dysfunctional ones (e.g., brooding). The tendency to savor positive aspects of the future was the main driver of anticipating less negative and more positive emotions. Our findings align with theories of aging suggesting that older adults report better well-being and better emotion regulation, and have implications for helping adults of all ages to manage their anticipated emotions when making decisions about the future.
Olsson, Henrik Co-authors: Wändi Bruine de Bruin; Mirta Galesic and Drazen Prelec
“Predicting the 2018 US House of Representatives elections based on social-circle expectations, state-expectations, and Bayesian Truth Serum”
E-mail: olsson@santafe.edu
The majority of election polls ask people which candidate they will vote for (own intentions). The accuracy of election predictions, however, can be improved over and above these own-intention questions by asking people who they expect will win the election (winner expectations; Graefe, 2014). A related approach recently studied by Galesic et al. (2018) is to ask people about the voting intentions of their close social contacts (social-circle expectations). These questions were better than own intentions and winner expectations in predicting two recent presidential elections in the US and France. Another, complementary approach is to assign more weight to more credible answers. The Bayesian Truth Serum is a scoring system that penalizes careless or dishonest answers and may therefore also serve as an index of respondent quality (Prelec, 2004). In a national longitudinal survey (N>4,000), we compared predictions of the 2018 US House of Representatives elections based on own intentions, state-winner expectations, social-circle expectations, and Bayesian Truth Serum based on either own intentions or social-circle expectations. Social-circle expectation questions outperformed own-intention questions at the national and state levels in predicting the vote-share distribution (in two waves) and the margin between Democrats and Republicans (in all three waves). They outperformed state-winner expectations in all three waves. Additional improvements were achieved when social-circle expectations were combined with the Bayesian Truth Serum. Social-circle expectations and the Bayesian Truth Serum show promise as an alternative to own-intention and winner-expectation questions.
Eriksson, Gabriella, Co-authors: Wändi Bruine de Bruin, John Maule and Nichel Gonzalez
“Drivers’ decision to speed: Perceived risk factors and reasons for and against speeding”
E-mail: g.eriksson@leeds.ac.uk
According to the World Health Organization, reducing drivers’ speed is crucial because it leads to fewer and less severe accidents. Here, we report on mental models interviews designed to characterize the factors drivers perceive to contribute to accidents, whether speeding is perceived to be one of those factors, and what drives their decisions to speed. The interviews were designed to reveal people’s beliefs or mental models about driving –as a first step towards identifying potential needs for intervention. Specifically, we used mental models interviews to investigate how drivers perceive accident risks and decisions to speed. Semi-structured ‘mental model’ interviews with 54 drivers who varied in experience were conducted in the UK and Sweden. Drivers were asked about risk factors in driving and discussed why they speed, as well as why they avoid speeding. Across interviewees, speeding was the second most identified risk factor after overtaking. The third most mentioned risk factor was mobile phones. Most interviewees considered speeding as going faster than the speed limit, but some only considered driving up to 40 miles per hour over the speed limit to be speeding. Most interviewees reported being in a hurry as a reason for speeding, followed by feeling safe in the car and enjoying the speed. Reasons against speeding included law enforcement, increased stopping distances and the potential risk of hitting a person. Thus, interviewees did identify speeding as a risk factor, but seemed to have some misconceptions of what speeding means. We discuss our findings in terms of their implications for risk communications relevant to driver education materials, dashboard tools and variable road signs.
Kause, Astrid
“Consumers’ perceptions of rules for reducing carbon footprints of food”
E-mail: a.kause@leeds.ac.uk
Consumers’ food choices substantially influence the carbon footprint associated with food production and consumption. Environmental and climate researchers have identified effective rules for reducing the carbon footprint of consumer food choices, such as eating seasonal produce and limiting consumption of red meat and dairy. In a national UK survey, we examined whether consumers were aware of the most effective rules for reducing the carbon footprint of food choices. On average, our participants knew 1-2 rules, which were relatively ineffective for reducing the carbon footprint of food choices. Even participants who self-identified as pro-environmental did not always generate the most effective rules. We also asked participants to estimate, either in grams of greenhouse gas emissions or percent change, how much carbon footprints would be reduced by implementing specific rules. Those who were asked to make estimates in grams (which is the prevalent unit for describing carbon footprints in environmental sciences) were less accurate than those who made estimates in percent change. Our findings suggest a need for interventions that help consumers to identify more effective rules for reducing the carbon footprints of food choices. Communicating these in percent change likely makes rules easier to understand, and should allow consumers who are motivated to reduce the carbon footprint of their food choices to do so more effectively.
Konstantinidis, Emmanouil, Co-authors: Maarten Speekenbrink, Leonardo Weiss-Cohenand David R. Shanks
“Expected Value Insight Learning: A computational model of confidence in experiential choice”
E-mail: em.konstantinidis@gmail.com
The relationship between decision accuracy and confidence has been extensively investigated since the early days of psychological research. Yet little is known about decision accuracy and confidence in consequential economic decisions as in the case of experience‐based decision‐making. Here we examine the progression of choice and confidence judgments in a typical experience‐based task and provide a formal description of confidence judgments and their relationship to choice mechanisms. Participants select cards from four different options with different payoffs. They are then asked to provide a confidence judgment on having selected an advantageous option. The results show that learning to choose the most profitable options is accompanied by accurate confidence reports. We develop a model to unfold the cognitive processes of choice and confidence based on Reinforcement‐Learning and Signal Detection Theory. Our model accounts for both choice and confidence and provides a framework of metacognitive judgments in the context of experience‐based decision‐making.
Okan, Yasmina, Co-author: Michelle McDowell
“Does iconicity really matter? Examining how the design of icon arrays affects key outcomes for health decisions”
E-mail: y.okan@leeds.ac.uk
Graphs are increasingly used and recommended to facilitate risk communication and support decision making. Icon arrays in particular can substantially improve risk understanding and reduce common judgment biases. Such displays can vary in their level of iconicity, ranging from more concrete, high-iconicity arrays (e.g., person-like icons) to more abstract, low-iconicity arrays (e.g., plain circles). Theoretically, high-iconicity arrays could facilitate recall of risk information owing to their pictorial nature and to their more direct links to the object they represent. However, previous work examining iconicity effects has yielded mixed results. We conductedthree experiments to shed light on this issue. Participants (n= 3,136 across experiments) were presented with icon arrays depicting the risk of being diagnosed with Type 2 diabetes or a hypothetical virus. The arrays contained either plain circles, smiley/frowny faces, or person-like icons. Experiment 1 also included a control numerical-only condition. We assessed participants’ numeracy and graph literacy, as there is some evidence that such cognitive abilities may moderate iconicity effects. Supporting previous work, we found that icon arrays improved risk recall relative to numerical information. However, we found no consistent main effects of iconicity or interactions with cognitive abilities for risk recall, risk perceptions, and affective reactions. Smileys were rated as more familiar than the other icon types, but often received more negative evaluations, particularly relative to person-like icons. Taken together, our results suggest that icon arrays are an effective method to improve risk recall, but their beneficial effect may be largely independent of iconicity.
Okan, Yasmina Co-authors: Eva Janssen, Mirta Galesic and Erika A. Waters
“Using the Short Graph Literacy scale to predict precursors of health behavior change”
E-mail: y.okan@leeds.ac.uk
Graphical displays can facilitate risk communication and promote better health choices. Their effectiveness in improving risk comprehension is influenced by graph literacy. However, the construct of graph literacy is still insufficiently understood. This is partly because existing objective measures of graph literacy are either too difficult or too long for the general population. We conducted secondary analyses to construct a new 4-item Short Graph Literacy (SGL) scale, which we adapted from an existing 13-item scale. The initial construction was based on data collected in a laboratory setting. We then validated the SGL scale using data from nationally representative samples in Germany and the US. We also examined how SGL scores relate to key cognitive, affective, and conative precursors of behavior change described in theories of health behavior. To this end, we performed secondary analyses of a third study involving a nationwide US sample comprised of 47% racial/ethnic minorities and 46% with limited formal education. Graph literacy was significantly associated with cognitive precursors in theoretically expected ways (e.g., positive associations with risk comprehension and response efficacy, and a negative association with cognitive risk perception). Results for affective precursors generally mirrored those for cognitive precursors, although numeracy was a stronger predictor than graph literacy for some affective factors (e.g., feelings of risk). Additionally, graph literacy (but not numeracy) predicted key conative precursors such as defensive processing. Our data suggest that the SGL scale is a fast and psychometrically valid method for measuring objective graph literacy. Our findings also highlight the theoretical and practical relevance of this construct.
Summers, Barbara, Co-author: John Maule,
“The effects of time pressure on managerial decision making”
E-mail: bs@lubs.leeds.ac.uk
Research in organizational psychology shows that people often feel there are insufficient hours in the day and that they have to work longer and harder than ever before (Menzies, 2005). However, there has been little interest in the effects of time pressure on key managerial activities such as decision making. This is surprising given that laboratory-based research shows that time pressure affects the processes and the outcomes of decision making, often in detrimental ways (Ordonez, Benson & Pittarello, 2015; Svenson & Maule, 1993). The objectives of this study are to investigate whether the strategies used to adapt to time pressure identified in laboratory research (e.g. filtering, acceleration) are used by experienced managers and, if so, the impact thesehave on decision outcomes and the emotional state of the decision maker.
Our findings confirm that strategies for coping with time pressure identified in laboratory research are used by experienced managers in the workplace and we identify two further strategies. We also reveal two different ways that time pressure is appraised (challenge and hindrance) each linked with a different emotional state and strategy choice.
Taylor, Andrea, Co-authors: Astrid Kause, Barbara Summers and Melanie Harrowsmith
“Exploring public responses to impact-based weather warnings in the UK”
E-mail: a.l.taylor@leeds.ac.uk
The UK Met Office issues impact-based severe weather warnings. Unlike earlier phenomena-based warnings, they are based on the local impacts of severe weather rather than meteorological conditions alone. To decrease risk from severe weather, it is importantto understand how members of the UK public interpret and act on these warnings. This paper addresses this through a post-event survey (n=552) conducted following Storm Doris, a 2017 winter storm during which wind warnings were issued across much of the UK. Survey questions examined 1) understanding of impact-based wind warnings; 2) predictors of perceived forecast trustworthiness; and 3) predictors of recalled and anticipated action. We find that UK residents generally understand that weather warnings are based on potential weather impacts. However, many do not realize that the conditions needed to trigger a warning in one area may be different to those in another. This could lead some to underestimate the risk implied by local weather warnings. Forecast trustworthiness is predicted by trust in the Met Office, perceived “understandability” and receiving a more severe warning. Recalled and anticipated protective responses are associated with prior experience of strong wind, level of concern evoke by the warning and interest in weather. Trust in the Met Office additionally predicts anticipated protective response. Differences in local warning level were not found to affect either recalled or intended behavior response. This study highlights the importance of institutional trust in the effective communication of severe weather warnings, and a need for education on impact-based weather warnings. Above all, it demonstrates the need for further exploration of the effect of weather warnings on protective behavior.
van der Wal, Natalie
“Risky Decisions and Risk Communications in Real-World Emergency Evacuations”
E-mail: c.n.vanderwal@leeds.ac.uk
Fires, terrorist attacks, and other threats require emergency evacuations. Faster evacuations during emergencies could save more lives. Policy makers therefore seek to improve emergency evacuation preparedness, in case of fires, terrorist attacks, and other threats. However, people may engage in behaviors that undermine the effectiveness of emergency evacuations, thus putting themselves and others at risk. We took risk research out of the lab to characterize risky decisions and risk communications in the context of real-world emergency evacuations. We collected 126 videos of real-world emergency evacuations during fires, terror attacks, and other threats. The videos of these emergency evacuations were filmed by members of the general public, CCTV cameras, tvnews crews, and others. We identified risky behaviors and risk communications and situational variables such as the visibility of the threat. The most common risky behavior pertained to slow initiation of evacuation due to for example, collecting items before evacuating and confusion about whether the emergency was real or not. We identified risk communications, such as the sounding of an alarm, having staff guide people out, and making a live or pre-recorded announcement. Our analyses suggested that the most effective risk communications for reducing risky evacuation behaviors included announcements or staff guiding people out. Especially when threats were not visible, alarms alone were often ineffective and had to be combined with either announcements or staff guiding people out. We will discuss our findings in light of research in risk perception and communication, and suggest improved guidelines for emergency communication that will ultimately save lives.