Search

In description-based decision problems, outcomes and their probabilities are clearly stated. People tend to be overly sensitive to rare events. However, in real life setting, decisions usually have to be made based on experience. People tend to be under sensitive to rare events (Hertwig and Erev, 2009). The existing reasons for this description-experience gap include small samples and recency effect (Hertwig and Erev, 2009). The current study aims at examining if memory biases can also be a mechanism that explains this description-experience gap. Madan et al. (2014) suggested that the tendency to better remember extreme outcomes could be one of memory biases. They found that extreme outcomes are overweighted in memory and this overweighting correlated with risk preference in decisions from experience. Madan et al. (2017) found that this memory bias correlated with risk preferences in decisions from experience, but not decision from description. However, these studies did not involve any rare events, but risky choices with equal probabilities. Thus, we change the saliency of rare events to manipulate memorability and then test its influence on risky choice. More specifically, we plan to change the visual saliency of rare events, such as colour and font size, to make them distinct from other outcomes (Kunar et al., 2017), because distinctive item is more likely to be recalled compared to other items that do not have distinctive feature (von Restorff, 1933).

This study aims to examine if people make risky choices differently depending on whether the information comes without a clearly identifiable source or with a clearly stated social source ('others'). This study will provide an insight into how people make risky choices and the factors that influence those choices. Long-term, we hope that these results will help improve our understanding of how people use information to make choices, perhaps leading to some new techniques to improve decision-making that needs to consider personal and social experience.

Our laboratory's Implementation of the Mazar_Srull RRR protocol

This project examines whether sensory preconditioning could be explained via the mental replay of previous training in periods of rest.

Pre-registration, materials, and data from Exp 2 in Context Paper

Materials and Data from Exp 1 in Paper

Previous work by Dolan et al. (2012), Durantini et al. (2006) and Cialdini (1993) has demonstrated that who delivers a message to the public makes a key difference on the ensuing behavioural response to the message. This phenomenon has been aptly named the Messenger Effect. The current research project seeks to explore whether trust in the messenger has an effect on people’s perceptions and beliefs in the message, specifically in the context of Honduras during the COVID-19 pandemic. Like many other countries in Latin America, Honduras suffers from decreased trust in its governmental institutions as evidenced by the most recent public opinion report issued by Vanderbilt University’s Latin American Public Opinion Project (LAPOP) for the years 2018/2019 (Zechmeister & Lupu, 2019). The report highlighted that in terms of how much the population deemed their country’s basic institutions and processes as trustworthy and respectable on a scale from 0 - 100, Honduras scored an average of 43.3, placing it as the country from those surveyed by the project with the third lowest trust in its institutions. This is of particular importance in COVID-19 times because the ensuing crisis generated by the disease has exposed tears in the country’s already fragile social fabric, with the government currently facing backlash over their response to the virus and corruption allegations. Based on this evidence, we believe that an effect will be observed on the population’s trust and adherence to prevention measures issued by the Honduran government during the pandemic. To test whether messenger effects could exist in this context, we propose an online survey experiment where participants will be separated into two groups, a control and a treatment group. The control group will be viewing a set of social media messages with COVID-19 preventive measures presumably issued by the World Health Organisation, which is an institution familiar to most Honduras and of neutral political standing. The treatment group will view the same messages, but as issued by the Honduran government. Participants will then be asked to rate how much they trust the institutions sending the message, how much they believe in the effectiveness of the preventive measure being prescribed, how often they think this institution tells the truth and how much they intend to follow the recommendation stated by the organisation. This will be followed by a set of demographic questions including age, gender, political views and ethnicity. It is expected that the treatment group who views the message as sent by the Honduran government will on average across messages display less trust in the institution, less adherence to the preventive measure, less belief in the measure, and they will believe this institution tells the truth less often. In terms of demographics, it is hypothesized that men will show more trust than women for both institutions in line with what other studies have found using trust and reciprocity games (Chaudhuri & Gangadharan, 2002). In terms of ethnicity, it is believed that those who identify with Indigenous and African minorities will have a lower level of trust due to the discrimination that these groups have historically suffered in the region (Hooker, 2005; Patrinos, 2000). For age groups, we believe young people (16-34) will be those who will show the lowest level of trust than any other age group, in line with observations that recent protests and unrest across the region have been driven by youth (Groshek & Bachmann, 2014; Valenzuela et al., 2012). In the case of political beliefs, subjects who identify as having right-leaning political ideology will show more trust than left-leaning participants, this is based on the findings that partisan beliefs and associations are linked to trust (Carlin & Love, 2013). The current government in Honduras is right-leaning, this should promote higher levels of trust and belief in people aligned with these views.

In this study, we want to examine if people would act differently and have a different tendency of anti-social behaviours when making decisions using a second language and their mother tongue for themselves and other people

This project examines how anagram solution could be facilitated by encouraging mental replays of unsolved trials using contextual cues between solution attempts.

Some of our previous research has investigated effects of providing house-edge information to gamblers, which specifies the proportion of all money bet that is retained by the gambling operator (Newall, Walasek & Ludvig, 2020). That research focused on self-reported measures rather than behavioral outcome measures and was restricted to online casino games. House-edge information, however, is relevant to all gambling games, including sports betting. This project will investigate the effects of providing house-edge information to soccer bettors, using incentivized payouts and realistic odds based on a round of English Premier League matches. Soccer is potentially a useful gambling format in which to disclose house-edge information, given that there can be large differences in the magnitude of the house-edge between different types of soccer bets, even for the same matches. We plan to use the final round of fixtures from the 19/20 Premier League season (26th July).

This project investigates the reasons behind why people choose to engage in conspiratorial thought. There has been much research outlining the welfare negatives resulting from conspiracy theory beliefs, from increased political aggression and extremism (Rottweiler & Gill, 2020), to a higher probability of engaging in petty crime (Uscinski & Parent, 2014, Jolley et al., 2019). In addition, correlational research has mapped the various links between conspiratorial thought and individual differences, such as examining the associations between psychosocial stress, anxiety, and belief in conspiracy theories (Swami et al., 2016) and charting the associations between the Big Five personality factors and conspiratorial beliefs (Swami et al., 2011). An interesting question therefore concerns why people hold conspiracy beliefs, despite these associated negative outcomes. In this project, we consider the possibility that conspiracy belief may also have positive welfare benefits, under specific conditions. A credible answer to this question may come from the realm of ambiguity intolerance. Considerable research has found that people have a low tolerance and aversion to ambiguity and uncertainty (Furnham & Ribchester, 1995). In accordance with this, people often try to escape ambiguous situations. In this project, we suppose that, when in the environment of ambiguity, conspiratorial beliefs act as a mechanism that reduces ambiguity and thus increases localised affect (welfare). To escape ambiguity, people tend to jump to the first convincing explanation we might find, and conspiracy theories can often fill this gap due to their alluring nature.

In 2015, a trade organization which represents some of the UK‘s largest gambling operators, began a high-profile responsible gambling campaign which is now recognized by at least 54% of all UK adults, and which continues to this day (The Senet Group, 2017). The campaign involves the warning label, “when the FUN stops, stop” (where the word ― “fun” is printed in noticeably larger font than any other word). Newall, Walasek, Singmann and Ludvig (2019) conducted an initial experimental test on that warning label’s effect on behavior (N = 506). They found a slight but insignificant increase in gambling when the warning label was shown, and concluded that the warning label was not effective in its stated goal of prompting more responsible gambling behavior. However, Newall et al’s conclusions were disputed by Gilmore (2019). Gilmore stated that the incentivization used (£0.90 per participant) presented insufficient risk of loss to participants. Gilmore also stated that the money risked came from an experimental bonus, rather than coming from the participant’s own funds (it is unethical for experimental gambling research to risk a participant’s funds outside of the experimental context). Next, Gilmore stated that the design of “when the FUN stops, stop” was an older design, and the image used omitted reference to GambleAware and the 18+ age restriction detail. Here we present an experiment designed to address these limitations. We have designed an experiment where participants are initially recruited to retype captcha codes. This is an example of a “real effort task” used in experimental economics, which is a way to minimize endowment effects from providing participants with an experimental bonus to risk from. Participants will then be presented with a fully-functioning and realistic online roulette game, where they can risk as much or as little of this bonus as they wish. Warning label presence will be manipulated between-participants, with the warning label graphic taken from the Senet Group’s current online brand guidelines. We also note that since Gilmore (2019), an alternative design of “when the funs stops, stop” has been seen by us in some industry marketing materials, noticeably where the entire logo is shown in an identical font and size on a grey background. The present study will therefore also test this newer warning label’s effect on behavior.

A gambling task and comprehension question to investigate whether the way a warning label is framed can effect people’s gambling behaviour, and comprehension of gambling odds.

We will replicate the core design from the experiment in Madan et al. (2017), but with different visual stimuli. This study further analyses property of the scenario permits the subjective identification of two distinct contexts which tries to answer this question. It also adds to the investigation of whether the context of encoding or the context of retrieval is crucial for determining which outcomes are overweighted in choice.

Rationale Previous research found that displaying variations of the “when the fun stops, stop” during an incentivized online gambling session had no effect on the proportion of available funds bet (Newall et al., 2022). These findings cast doubt on the efficacy of that particular safer gambling message. In recent months the UK’s Betting and Gaming Council unveiled a new safer gambling message intended to replace when the fun stops, stop: “take time to think”. The present work will perform a similar evaluation of the take time to think message with the online roulette paradigm from the previous work, with a few design improvements. The primary goal of the current work is to evaluate whether the “Take Time to Think” safer gambling message leads to a change in concurrent gambling behaviour on an online roulette task. We will compare the proportion of funds that people bet when presented with either no message or the “Take Time to Think” message in two formats: as a banner or as a pop-up message. In the previous study (Newall et al., 2022), although the safer gambling messages were displayed prominently on screen during the gambling task’s instructions and during the task itself, participants may not have been sufficiently aware of the message for it to affect their behavior. Previous research suggests that “pop-up” messages, displayed in the center of the screen and which must be clicked in order to proceed with gambling, are the most effective at shaping behavior (Monaghan, 2009). Therefore, the present research will have three conditions: (i) no message (control condition), (ii) message only on-screen, and (iii) pop-up on initiation of the gambling session, plus message on-screen throughout. In addition, safer gambling messages might have no immediate effect on gambling behavior, but might have other influences on gamblers. To assess such other impacts, the present work will provide participants with an informational page directly after the gambling task, which gives an overview of the support provider GamCare’s services and a button that can be clicked that will take participants to the GamCare website upon completion of the study. Rates of clicking on that button will be compared between groups. In addition to the proportion bet or participants' interest in visiting the GamCare website, the take time to think message might also affect gamblers speed of play. Therefore, the present research will also test the mean speed of play between any two spins on the roulette wheel in the two gambling message conditions compared to the no-message control condition. Finally, any change in speed of play may be offset by changes in the total session time. To better assess this possibility, we will also examine the total number of roulette spins made by participants.

This study aims to investigate how risk-taking behaviour changes when people coordinate.

In our study, we want to test whether people will choose differently for themselves and others under ambiguity. In this experiment, participants will choose between two options with differing ambiguity levels both for themselves and for others. Moreover, we are interested in whether people will choose let a computer decide due to blame aversion

This is a pre-data collection pre-registration before Exp 3b. This experiment is the 7th in a series in this project. The pre-registration document for this experiment is in the pre-registration folder.

Soccer warning labels

This project investigates how social information and experience influences decision-making for self and others.