Executive MSc Behavioural Science teaching image_banner

Executive MSc Behavioural Science Dissertation Abstracts

Our Executive MSc Behavioural Science programme is a unique and dynamic programme for full-time professionals which aims to uncover the science behind behaviour. 

On this page you can read selected abstracts from the dissertation research projects conducted at the end of the programme, separated by overall theme.


If I Empathize, Will You Compromise? Exploring the Effects of Empathic Messaging on Behaviour around Impositions

By Cortney Price

The subject of empathy triggers strong opinions and interest in both scientific and non-scientific literature. There are those who are for empathy and those who are against, but most seem to agree that empathy is powerful. There seems to be a wealth of experiments which attempts to introduce a manipulation that fosters empathy within subjects for someone else in order to promote a behavioural effect. Examples include fostering empathy with the less fortunate to motivate charitable giving (Andreoni, Rao, Andreoni, & Rao, 2011) or boosting empathy in medical students to improve bedside manner and patient trust (MacLean, Kelly, Geddes, & Della, 2017; Shapiro, 2008). However, the literature seems wanting for manipulations that study the reverse scenario. Instead, this study hopes to foster the feeling of ‘being emphasized with’. If someone feels empathized with, are they more likely to be open to suggestion? Do they view a subsequent imposition with less reticence? Can they be persuaded to take actions that go against their immediate economic interests? In online crowdsourcing, time is money. Microtask crowdworking platforms represent the main source of income for a surprising proportion of crowdworkers, and many of these individuals are close to the poverty line (Gleibs, 2017). Today’s “gig economy” may not be as empowering as initially projected. Despite its non-classical setup, crowdworking still creates an employer-employee relationship between requester and crowdworker (Martin et al., 2017). This scenario seems to lend itself well to an experiment involving empathic messaging and economic imposition.

Through an experiment, I will explore the potential effects of the feeling of “being empathized with” in the context of online crowdworking. I will attempt to understand the effects of an empathic message – both cognitive and emotional – on crowdworker behaviour in the context of a subsequent ‘ask’ or imposition. I hope this research may shed light on ways to improve the requester-crowdworker relationship to benefit online research. I also hope to take a small step toward providing insights into how to craft more effective messages to be used in various communication scenarios, including the communication around public health risks and the recommendation of risk-reductive behaviours.



Consumer behaviour

The More Choice, the Better? Or Choice Overload? The Impact of the Number of Products on a Printed Catalogue’s Sales Count

By Dirk Görtz

Although choice in general is good, sometimes it appears that many choices can become too many. This choice overload can be a problem for sellers and buyers, as at worst such an overloading might end up in no choice at all. With dozens or sometimes hundreds of products shown, the printed catalogue is a medium with the purpose of offering choice. Up to now, no research has analysed the impact of more choice on a catalogue’s performance –a research gap that the present work considers.

The research aimed at answering the research question if there is any difference in sales count (conversion rate and average shopping basket) between a printed catalogue containing relatively more products and one that contains relatively fewer products. This was done with the help of an A/B testing in a block-randomised between-subjects natural field experiment. The data analysis revealed a statistically non-significant treatment effect on the dependent variables conversion rate (+11.7%) and average shopping basket (+2.2%). Besides, there were statistically significant treatment effects on the number of products sold (1) regarding certain product categories and(2)when comparing products displayed in the catalogue with those only displayed in the web shop.

The research made several theoretical contributions: Firstly, although the positive treatment effect on conversion rate and average basket was non-significant, these results suggest that there is no risk of overloading recipients. Secondly, the findings open up a new research field worth further investigation, namely the impact of ‘what is shown’ in catalogue on its performance: (1) If there are product categories to which consumers react more significantly if given more choice, this might open up new strategies for designing bigger catalogues. (2) Besides, if showing more products can lead to a sales increase of those same products, this might lead to new strategies for targeted sales.

As implied above, this study has not closed the research gap but rather has opened up a new research focus: showing more of the right products could be a way to achieve statistically significant and economically worthwhile performance improvement when increasing a printed catalogue’s number of products. Therefore, the underlying research has to be multiplied, extended and evaluated against the background of COVID-19.



The Return: How Applied Theatre & Behavioural Science Influence Acceptance of Interpersonal Discrimination among Female Students in Armenia

By Natalia Voutova

This research examines how the combination of applied theatre and behavioural interventions (receiving a letter from a psychologist/wring a letter to your future self (as a commitment contract) could reduce the self-reported level of acceptance of some forms of discrimination (gender based, against people with disabilities, against people with low socioeconomic status) among female university students in Armenia.

The experiment evidenced a statistically significant effect through the writing of a letter to one’s future self, after the students watched a play of applied theatre about social problems in Armenia while the other intervention (receiving a letter from a psychologist) did not produce a significant effect. There were no significant difference established between the main effect of receiving a letter from a psychologist or writing a letter to one’s future self, compared to the other treatment on the reduction of the self-reported level of acceptance of interpersonal discrimination, but nonetheless writing a letter to your future self is better. The effect was measured with a pre and post identical questionnaire comprising 13 vignettes anchored questions. The whole experiment was carried out in Armenian language.

The current study is highly relevant to the process of further development and diversification of behavioural interventions in non-WEIRD i.e., Western, Educated, Industrialized, Rich, and Democratic context.



Unshackling Learning from the 19th-Century Factory Model of Education: Mobile Game-Based Learning for Adult Distance Learners

By Elizabeth Close

Video game use continues to explode while educators have long sought to harness games to increase learning outcomes in educational environments. While much research has investigated the power game-based learning to engage, interest, and motivation students and children, much less has been done to examine whether games can increase interest among adults in the general population. Even less research, whether on students or adults, has been done on distance learners who learn ‘in the wild’, and even less on mobile phone learning. This study investigates whether mobile game-based learning (mGBL) can increase interest in learning to code in adults. Further, this study examines whether mGBL can increase what Dweck (2006) termed ‘a growth mindset’, which may motivate further learning interest in subjects that may seem intimidating or methodologically foreign. To compare traditional instruction versus a gamified approach, 90 participants participated in a randomized control experiment. One group learned coding with Google’s Grasshopper through mGBL, while the controls viewed the same content on slides delivered on smartphones. The findings showed that there was a positive relationship between age and interest in further learning to code after participation in both mGBL and text-based learning modules. Contrary to expectations, while both groups showed significant increases in interest and an adoption of a growth mindsets, there was no significant difference between the game-based or text-based groups. This study finds that there is promise for changing the domain of mindsets in adult populations, which may inform active learning curriculums and mGBL interventions targeted at older workers to improve STEM learning outcomes in the context of an ever-evolving automated world.

Can Loss Aversion Increase Effort and Performance in Education?

By Guy Dickinson

In recent years there has been a democratisation of access to online education for adult learners via ‘massive open online courses’ (MOOCs) which have provided access to university calibre education for free or very low cost (King et al. 2014). While MOOCs offer a rigorous and rewarding educational experience, the ‘drop out’ rate is typically over 80% for the majority of courses (Kizilcec et al. 2013), (Crossley et al. 2016). This high non-completion rate is often attributed to sharp decline in commitment as the course progresses (Clow 2013).

“Most education policies will fail if students do not exert effort. Yet, surprisingly little is known about what motivates students to invest effort in school or the causal impact of this effort on learning and achievement” (Levitt et al. 2013) p.28

What if adult learners could be motivated to increase completion of MOOC activities using a method that is low cost, and easy to implement at a scale that would suit the context of online education? I believe a behavioural change in adult learner motivation would provide a profound opportunity to increase access to education. My objective for this paper is to understand if such a change is possible.

In the last five years, behavioural scientists including Levitt have closed this gap in the literature, identifying ways in which incentives, and in particular, loss-framed incentives, have meaningful effects on learners' motivation.

I have developed a set of research questions that build from this literature, and analyse whether the positive affect of incentives on learner motivation observed in experiments with schoolchildren and college age students can be delivered within a single assessment experience such as those found in MOOCs, and whether these incentive effects crossover to adults.

To understand whether design changes in a MOOC can increase motivation in adults, I designed and ran a randomised control experiment with 493 adult learners, comparing the incentive effects of financial and non-financial loss-framed rewards on a common education activity — completing an assessment.

I found that for a certain segment of the experiment's participants, there was a statically significant effect from the pre-awarding of a 100% score — this non-financial incentive created a loss aversion response that increased effort in participants by a significant margin compared to the control. I also discovered that paying for completion of an educational task inadvertently created a marketisation of perceived effort, resulting in significant demotivation for a different segment of participants.

I use these findings to discuss implications for designing a cost-effective and scalable behavioural intervention that could increase motivation for adult learners, and reducethe precipitous drop out rate currently affecting MOOCs.


Finance, banking and investment

We Know a Consumer’s Financial Profile Predicts the Probability of Financial Distress, but What about Their Personal Traits?

By Juskurran Hothi

Many individuals from lower socioeconomic background are denied affordable credit as they have a very thin credit file and therefore perform poorly in existing credit scoring methodologies. This paper explores and extends the current literature on alternative inputs for scoring methodologies and investigates the ability of non-financial factors to predict financial distress. The results, from a LASSO regression on a representative sample of UK Consumers, show that markers for mood, time and risk preference do not meaningfully predict financial distress, over and above traditional credit markers.

Measuring the Effects of Behavioural Biases and Personality Types on Early Stage Investment Decisions

By Jose Miguel Smith

Despite there being a large amount of research dedicated to identifying and assessing the effects of behavioural biases in investment decision-making, the overall findings have been broadly inconclusive. The purpose of this study is to determine the extent to which behavioural biases and personality types influence early stage investment decision-making. It attempts to isolate and measure the effects of the Stereotype Bias, the Availability Heuristic and the Herding Effect by reducing the number of variables at stake during an investment decision via experimental design. Selected investor personality traits (Big-5 and Need for Cognition) are analysed as potential moderators of the investment decision. 801 investors were recruited to participate in an online randomised controlled experiment where their willingness to invest in a hypothetical early stage investment in a private company was measured. While the study finds a null effect for all three behavioural biases under review, it finds a significant effect for the Availability Heuristic by excluding investors that show a strong positive preference for the investment and controlling for Age. In addition, Age is found to significantly influence the early stage investment decisions of investors with low to medium levels of Need for Cognition. The study serves as an important first step in deconstructing the investment decision process into individual components. It provides previously unavailable baseline information and insights to form part of subsequent experimentation in investment decision-making.

Deposing the Disposition Effect: Experimental Evidence from Professional Traders

By Benno Guenther

Various types of investors exhibit a strong preference to sell stocks that have increased in value since purchased relative to stocks that have decreased in value. Shefrin and Statman introduced this phenomenon in 1985 and termed it the disposition effect. The disposition effect is a puzzling anomaly in finance and frequently associated with lower investment returns. One of the leading explanations for this anomaly is a combination of prospect theory and mental accounting. Conducting a quasi-experiment this study examines the disposition effect in the context of professional traders. Contrary to previous studies this study does not find evidence for the existence of the disposition effect for professional traders. On the other hand, this study finds evidence of the reverse disposition effect after a simple intervention. Moreover, the study examines the disposition effect in the context of long as well as short positions but no evidence is found that the magnitude of the disposition effect differs between long and short positions. While the disposition effect is often described as detrimental to investment returns, the results of this study suggest that a higher disposition effect is associated with higher trading performance. However, this unusual finding can be attributed to the mean-reversion property of the underlying security data set.

Does a Prompted Consideration of Remaining Life Expectancy, Together with Hopes and Fears for the Future, Change the Investment Decisions of People about to Retire?

By Philip Broadley

"UK Government policy now allows people the freedom to choose how to use their accumulated pensions savings at retirement.  Evidence shows that people generally are loss averse and under-estimate their remaining life expectancy at retirement.  If these biases are applied to individuals’ decisions about their use of pensions savings, then they are likely to achieve sub-optimal returns.  They will view their future with too short a time horizon, allocate too much of their capital to low risk and low return assets: thereby reducing the income available to them in retirement.  The consequences of sub-optimal investment are particularly important for those people who retire with below median levels of pension savings.  Each 1% of investment income forgone typically reduces their annual income in retirement by 3%.

This paper explores whether an appeal to the emotions of experimental subjects making a hypothetical investment decision changes their behaviour.  The results of a randomised control trial (RCT) provides some evidence that such an emotional appeal has an effect on asset allocation, as it gives rise to more thoughtful consideration of investment choices.  The RCT also points to a consistent underestimation of remaining life expectancy at retirement.  It also shows that people generally share the same hopes and fears about their retirement.

The paper discusses how these findings could be applied to the advice processes offered to those reaching retirement, particularly those who cannot afford independent financial advice.  It suggests that appealing to emotions can help people make an infrequent and complex decision."

Does Having a Wider Perspective Help Reduce Dishonesty in the Banking Industry?

By Kyo-Hwa Kim

This paper investigates why bankers misbehave, and what can be done to get them to cheat less. By replicating a previous study, this paper shows that investment bankers based in the U.K. cheat more when reminded of their professional identities. What is interesting is that they do not show signs of excessive cheating unless they are reminded of their jobs. A list of reasons people cheat, such as ego‐depletion, lack of reciprocity, social and altruistic cheating, extrinsic motivation, fixed mindset organization culture, moral numbing in markets, lack of selfawareness, informal nature of training, short termism, and stereotype internalisation area discussed. The paper goes further to look at ways to reduce cheating even when bankers are primed of their professional identities. By reminding bankers of their life outside of work and the longevity of their careers, it is possible to reduce cheating behaviour, while keeping work identity priming is still effective.

The observation that bankers cheat excessively only when reminded of their jobs means that the common stereotype of bankers as greedy and dishonest people, who self‐select themselves into the banking industry is unfounded. It also implies that the work environment of bankers induces them to cheat. Lastly, this paper shows there is a simple and cheap way of reducing such cheating behaviour, while at the same time improving employee life‐satisfaction. Ever since the financial crisis of 2008, banking supervision has been skewed towards more regulation, controls, and increased personal criminal liabilities. The findings of this paper open up the debate for an alternative route towards investment banking culture change.

Swearing Bankers – Can an Oath Influence Behaviour in the Financial Services?

By David Grosse

Concern over the behaviour of bankers intensified following the financial crisis of 2008 and yet, despite large amounts of focus from governments, regulators and the finance industry, evidence of unethical conduct continues. With the traditional levers of sanction, legislation and regulation having limited impact, attention has turned to other approaches that may influence banking culture and behaviour. One instrument that has been introduced in both the Netherlands and Australia is the Banker’s Oath, albeit with little evidence on effectiveness. This dissertation examines the potential impact of oaths on bankers’ behaviour, reviewing existing literature and knowledge on their efficacy, in particular with reference to the role of cognitive dissonance and commitment devices. A randomised control trail was also undertaken on 292 finance professionals covering cohorts recruited through Prolific Academic, LinkedIn and directly via employer banks. All participants answered a survey across a range of ethical scenarios, assessing how likely they were to take advantage of a situation. Participants were allocated into two conditions, one of which commenced the exercise with exposure to a Banker’s Oath, the other included the same wording after the ethical scenarios had been concluded.

The key hypothesis was that prior exposure to the oath would reduce the amount that people would answer in a way that indicated they would exploit a scenario, and that the dependent variable, of a combined ethicality score, would be impacted. Whilst the treatment scores across all cohorts reduced, the results did not show a statistically significant effect.

However, the nature of the oath within the experiment was light touch, in that it simply involved reading the relevant framing question and did not require explicit personal commitment, swearing nor any accompanying ceremony. The majority of participants acknowledged that if asked to take a Banker’s Oath it would have some impact on their behaviour at work, and as such further research with a more realistic and resonant oath setting would be instructional.

A Study of the Impact of Information Overload on Financial Market Decision Makers

By Mark Schofield

"It has been estimated that ninety percent of the world’s data has been created in the past two years. Information now comes at us faster than we can process it. New technologies have evolved to distribute and consume information faster, but our neural processes are not able to keep pace. Information overload is the result. This dissertation examines the effect of information overload on decision-making among finance professionals and how experience may shape this. Effective decision-making under pressure is critical to the roles of financial market participants and for the stability of the financial system.

A randomised control trial was conducted on 412 financial market professionals, including traders, brokers, asset managers and analysts. Participants were given different amounts of information to absorb and then asked to carry out two tasks. The first was an estimation task that asked them to predict the behaviour of the group, simulating a market environment. The second was a recall task, to test accuracy. Two hypotheses were tested. The first was that information overload would cause decision makers to adopt an intuitive, rather than analytical, decision style. This could leave them exposed to cognitive biases that could adversely affect their decisions. The second was that experienced decision makers would be less affected. The study shows that inexperienced decision makers switched from an analytical process to an intuitive decision style when information overload was induced. Experienced participants displayed more consistency in their decision style, but they displayed an intuitive decision style in both the control and treated conditions.

The results show that performance in the estimation game deteriorated when decision makers used an intuitive, rather than analytical, decision style and that accuracy in the recall task deteriorated when information overload was present. If decision makers are aware that information overload can affect decision style, and that intuitive decision making can be ineffective when decision tasks are unfamiliar, they can take steps to manage these risks."




Are There Gender Differences in the Propensity to Compete among Chinese Individuals: An Empirical Investigation

By Gerald Wu

This empirical investigation explores gender differences in the propensity to compete among Chinese individuals. The study uses an online survey distributed to students in a university located in Shanghai and measures performance among Chinese men and women under different incentive schemes. While the majority of empirical studies to date have been conducted in Western countries, this research is the first to explore differences in competitive behaviour among Chinese men and women. Contrary to current findings from Western countries, the results of this study suggest that there are no differences in performance under competitive conditions between Chinese men and women. Furthermore, competitive incentives have an adverse effect on performance among Chinese individuals. Two main conclusions can be drawn from this study. First, the gender gap in senior positions within China is not attributed to gender differences in the propensity to compete. Second, cultural factors play a predominant role in nurturing behaviours. Therefore, policymakers must account for cultural differences in developing policies.

Turning a Unicorn Into a Horse: Increasing the Participation of Young Women in Computer Science by Making Communal Goals Salient

By Polly Lagana

Known as “unicorns” in the technology world, female computer scientists are as rare and highly sought-after as the mythical creature. Companies around the globe desperately try to recruit “unicorns” by throwing money at the problem, pouring significant resources into pipeline training programs as early as primary school to target first-generation, minority young women. Yet these skilled students, many shown to outperform boys in STEM, are not opting into technology majors and careers, especially at the critical points of entrance to university and entry-level employment.

Research shows that a lack of communal goals may explain low participation rates for females in computer science, yet few studies focus specifically on first-generation, minority young women. This study examines ways to propel the growing population of underserved girls with computer science experience into technology opportunities at the university and entry-level career level, with an emphasis on the variance in communal goals.

To test if highlighting “social impact” or “collaboration and community” increases participation, five randomized controlled field experiments were conducted in public high schools in New York City with 227 low-income, first-generation, minority young women (ages 17-19), many with computer science experience, who are preparing to enter university. At each school, students were invited to immediately apply to one of three randomly-allocated versions of the same university technology program: the original version or one of two treatment versions highlighting collaboration and social impact.

The results revealed higher odds of first-generation young women with computer science experience applying to the collaborative version of the program and lower odds of students with computer science experience applying to the prosocial version. The results were statistically significant although the small population means more research is necessary to gauge full effects.

The work suggests that educators and employers recruiting “unicorns” should consider making the collaborative elements of technology opportunities salient to increase participation in technology majors and job opportunities. Further research and analyzation around developing community and collaborative programming in technology is encouraged to ensure the new wave of underrepresented young women with computer science experience join the field.



Health and wellbeing

Impact of Meditation on Subjective Temporal Discount Rates of Monetary Rewards

By Madeline Quinlan

The recent explosion of research on mindfulness and meditation has spanned numerous fieldsand touted a variety of benefits to practitioners. There is less literature, however, investigatingimpacts of meditation on cognitive biases, specifically temporal discounting – the tendency toplace higher value on immediately available rewards – resulting in lower self-control andhigher current consumption. This work looks to determine whether a brief mindfulnessmeditation can significantly decrease subjective temporal discount rates.

Previous research provides support for four mechanisms common to meditation and temporaldiscounting: 1) activation of common brain regions 2) temporal distance effects 3) resistanceto ego depletion and 4) reduction of automaticity in judgement. These, taken within the contextof mixed reviews and criticism of certain meditation literature and a call for better rigor ofstudy, shaped the experimental design herein. 249 participants took part in a randomized control experiment at the London School of Economics and Political Science Behavioural Lab for Teaching and Research (LSE), randomized into three listening conditions: a podcast control recording, a jazz music track, and a guided meditation. Participants completed Kirby’s Delayed Discounting Task (DDT) to calculate implicit temporal discount rates, as well as Mindfulness Attention Awareness Scales (MAAS) to measure both state- and trait-mindfulness.

Contrary to expectation, those participants in the meditation treatment group exhibitedsignificantly higher temporal discount rates and present bias. Further, trait-mindfulness did notpredict lower discount rates exhibited by participants. I present two alternative explanations for this thought-provoking result: temporal focus & priming effects and backfiring of ego depletion resistance. Overall, this significant finding calls for further investigation into therelationship between mindfulness, meditation, and temporal discounting, ensuring thatpotentially backfiring meditation interventions are not used in the field.

The Effect of Behavioural Science Tools on Vaccination Uptake in Ukraine: A Randomised Controlled Trial

By Anastasia Nurzhynska

According to the World Health Organisation, vaccine hesitancy is among the ten greatest threats to global health. In Ukraine, less than 80% children are vaccinated, with one of the lowest levels of trust in vaccination and the highest number of measles cases in Europe. The negative effects of the anti-vaccine attitudes extend from health threats to the individuals and society, to the economic burden of treating patients with infectious diseases.

The situation is mainly caused by the anti-vaccination sentiments of parents. It is possible to taxonomise anti-vaccination positions according to specific biases. Several of the messaging techniques used by public authorities currently, which aim to raise awareness of the benefits of vaccination, have led to negative public responses. It is proposed here that behavioural science through messaging may influence parental decision-making and, in turn, may improve public health outcomes.

This paper presents the findings of a randomised controlled experiment (N = 738) investigating whether messaging adjusted according to behavioural science principles in terms of social norms, loss-framing, simplification, and messenger effects reduce mothers’ hesitancy towards the vaccination of their children. The study was conducted as a two-wave online panel. Mothers with one or more children were randomly assigned to receive one of six letters. The research examined the differences between mothers’ vaccination attitudes, intentions and behaviour depending on their exposure to different types of messages.

The results showed that the messages from doctor around social norms positively affect mothers’ vaccination behaviours. We conclude that vaccination mandatory messages used by the public authorities today are unlikely to reduce vaccine hesitancy and that parental demand for vaccination could be gained with the help of behavioural tools. We discuss the limitations of the current study, followed by a series of recommendations intended to inform further research and current practice.

Lung Cancer: An Avenue to Improve Referral Decision-Making in Primary Care

By Enrique Burton

Early diagnosis is the single most important predictor of a positive prognosis for anyone suffering from lung cancer (LC). Nevertheless, only a fraction of patients are diagnosed at an early stage, thus compromising their chances of survival. Indeed, LC is not only the most common cancer worldwide, but it accounts for nearly 1 in 5 cancer-related deaths. In primary care, the challenge of promptly referring patients for further diagnostic investigation is exacerbated by potential confounding from symptoms of other illnesses, possibly co-existing, that are very common in general practice. Therefore, this study aimed to investigate the potential for positively influencing primary care physicians’(PCPs) decision to appropriately refer potential LC patients. By employing a signal detection theory (SDT) approach, the study primarily evaluated two hypotheses: first, that an intervention based on the presentation of a system-generated differential diagnosis improves PCPs’ discrimination (d’) performance (i.e., ability to distinguish patients that should be referred); second, that the intervention reduces PCPs’ response bias (c), i.e., increases PCPs’ willingness to refer patients for urgent investigation.

To that end, the study leveraged and extended recent research to design an online, between-subjects randomized control trial, with an adequately powered representative sample consisting of 205 medical students. The experiment consisted of ten hypothetical clinical vignettes that presented cases that should (or should not) be referred based on symptoms and risk profiles. Participants’ differential diagnoses and referral decisions allowed the estimation of discrimination and response bias SDT parameters; in turn, linear regressions were conducted to evaluate potential average treatment effects (ATE).

Both hypotheses were confirmed: ATEs from the intervention were significant and yielded an effect size on discrimination and response bias equivalent to a Cohen’s d of 0.44 and 0.74, respectively. In addition, baseline results were consistent with previous research, providing confirmatory evidence regarding PCPs’ modest discrimination and bias toward not referring patients urgently.

The findings suggest that the implementation of clinical decision support systems that provide relevant diagnostic suggestions can significantly improve LC referral decision outcomes in primary care, thus providing potentially valuable insights for the consideration of healthcare professionals and policymakers.   

Framing Superbugs - Does Evidence Beat Practice? Testing the Efficacy of Advocacy Frames - An Online Randomised Control Experiment

By Dan Metcalfe

Antimicrobial resistance (AMR) is a significant threat to global health, requiring multifaceted policy action. Advocates for action must persuade others to act, and effectively framing AMR could plausibly increase support for change. However, recommendations about optimal communication to date have been developed using a traditional market research process, incorporating neither systematic testing nor the application of evidence from the behavioural science literature. This paper reports a randomised controlled experiment with over 1,900 respondents in the UK and the USA to test the impact of four treatment frames, derived using different methodologies, upon attitudinal and behavioural measures. The experiment was conducted with the cooperation of the Wellcome Trust, a UK-based health research foundation which works on tackling AMR.

The experiment finds robust framing effects of two out of four treatments, suggesting these findings have value in informing advocates of AMR about optimal approaches to driving change. Contrary to expectations, a frame derived through a traditional market research process improved behavioural response, as did a frame derived through a much faster and simpler application of evidence from the behavioural science literature. These findings have value in supporting advocates AMR advocates to increase the effectiveness of their communications. Furthermore, the study demonstrates potentially cost-effective ways to change or augment current market research practice with insights from behavioural science and the value of randomised studies to empirically test the impact of frames upon behavior. The study also indicates the potential for a more extensive programme of research.

Human-Centred Nudging: Improving Dietary Choices with Augmented Reality

By Elliot Harris

Obesity and nutritionally imbalanced diets are a rising concern in developed countries, and interventions are being designed to influence the dietary choices people make. However, the ecology in which purchasing and consumption decisions are made is complex. These complexities, plus individual dispositions and preferences, challenge the efficacy of any intervention that targets large groups of people at once. This paper explores the possibility of using personal, context-aware, augmented reality technology to overcome these challenges. A lab-based augmented reality experiment to influence dietary choices was subsequently conducted, using arrow cues based on approach-avoidance and attention mechanisms. The results reveal a small though insignificant effect of the intervention, and possible explanations and limitations are discussed. However, technological advances might afford a greater impact on behavioural interventions, be they dietary or otherwise, and hence may warrant further research.



Brexit Behaviourally: What Behavioural Lessons Can UK Government Communicators Learn from the 2016 EU Referendum?

By Tessa Buchanan

This study looks at how behavioural factors may have influenced the 2016 EU referendum. An experiment involving over 450 Leave voters suggests behavioural science might help to change attitudes towards immigration, and a survey backed by contemporaneous data illustrates MINDSPACE in action.

The experiment - a randomised controlled trial (RCT) – finds that a behaviourally-based narrative employing Affect is almost twice as effective as a fact at persuading Leave voters to donate to a pro-migrant charity (50% v. 28%).

The survey of Leave voters found:

  • Boris Johnson was a better Messenger (more likeable, similar and authoritative) than David Cameron; “experts” like the IMF were not recognised, and a “local businessman” was rated more trustworthy than a minister. BBC television and Facebook are identified as primary channels for this audience.
  • Incentives: ‘Take Back Control’ was seen as 4x more effective than ‘Take Control’. Most thought £4,300 per household per year was less than £350m per week (it is 6x greater).
  • Norms: Leave supporters were more likely to vote like family and friends than colleagues.
  • Default: The ‘status quo bias’ didn’t prevail. Were Leavers content and how did they see the risks? Results show low levels of knowledge about the status quo - most weren’t confident with terminology like “the single market”.
  • Salience may have influenced Leave voters who over-estimated the number of migrants in the UK.
  • Priming: Sub-conscious cues may explain why 34% thought the red Vote Leave bus represented Labour’s position and 12% thought a Hollywood poster represented UKIP’s.
  • Commitment/Consistency: Despite public statements to the contrary, most thought Cameron decided how to vote over a year before the referendum.
  • Ego effects may explain the “sovereignty spike”. Immigration now 3x more important.

In conclusion, Government communicators are encouraged to apply behavioural theory when developing campaigns and to adopt a ‘Test, Learn Adapt’ approach to establish What Works.

Priming Situational Attribution to Increase Poverty-Reduction Behaviour

By Malcolm Wong

Poverty and income inequality continue to be a key global concern, and this is mirrored in Singapore, which has seen a surge of interest in the uplifting of lowwage earners. Although much of poverty- reduction work is done at the macro institutional level, it clearly also requires social and political support at the individual level. To further our understanding of how to impact this, the current study investigates the potential for influencing an individual’s support for poverty-reduction through a behavioural prime to activate causal attribution for poverty. In an online controlled experiment, young adult Singaporeans aged 18-35 years old (n=213) were exposed to the treatment of a writing task prime that sought to activate a situational attribution of poverty, or alternatively, to a control writing task. Participants were then assessed on their willingness to engage in actions that address poverty, via an option to donate their potential lucky draw winnings to a charity focused on poverty. 

The results showed that the priming intervention did lead to thinking about situational causes for poverty, but it did not, however, correlate with higher donation amounts to the charity. Instead, a correlation between donation amounts and the education level of the participation was found. In interpreting the study, the importance of contextual factors in the effectiveness of behavioural priming is discussed with relation to pre-existing knowledge and concepts, participant education level, and the impact of online platforms; with the aim of advancing a more complex understanding of its limits and, more constructively, the factors that might enhance the effectiveness of behaviour primes.

Cheating in the Clear and Reducing it by Bringing Attention Back to One’s Moral Standards

By Raymond Au

An online experiment was conducted to study people’s propensity to cheat if they were given an opportunity to do so, even when the participants were asked to reveal their identities instead of remaining anonymous. It was found that people display a high propensity to cheat whenever an opportunity presents itself, but they do not cheat to the extent of obtaining the maximum possible gains. Moreover, at the point of temptation, people’s attention to moral standards was found to drop. We further showed that a simple intervention to bring people’s attention back to their moral standards will help to reduce the degree of cheating. While this study did not test whether people’s degree of cheating is determined by one’s amount of anonymity, past studies had shown that significant dishonest behaviour was observed when similar experiments were conducted under the context of participants’ anonymity. Therefore, we could deduce that anonymity promotes cheating. The surprising result from this study was that people will cheat even under clear identity, and this has major implications for policy makers.

Cultural Effects on Automatic Risk Responses

By Sulav Saha

Responding to risk is inherent to daily life and how we respond is arguably shaped by our social interactions, environment and values that represent the culture with which we associate. Social policy, marketing and messaging is often designed to influence our risk based decisions on predicted outcomes based on mainstream cultures. The trends in worldwide migration and growth in multicultural communities however increases the potential for unintended outcomes amongst minorities. The present research identifies differences in responding to risk between cultures through a matched-pair experimental design representing significant UK ethnic groups including mainstream White-British, South-Asian and South-East-Asian cultures. The results show a response to maximise opportunity when faced with risk with South-East-Asians taking 24% (p< 0.05) greater risk under ambiguity and 21% (p< 0.05) greater likelihood to risk losses for gain when compared to White-British. In comparison South-Asians take on 27% (p< 0.05) less risk under ambiguity and have 19% (p< 0.01) stronger emotive responses during risk events when compared to White-British. Further exploratory analysis is provided on British-Multicultural participants showing participants reducing their risk taking under ambiguity by 3% for every 10% (p< 0.01) of their lifetime spent in the UK. These results are linked to distinct cultural attributes and broader research that explain these differences.

Accountability In the Humanitarian Sector: A Tool for Better Humanitarian Action? An Integrated Review of the Role of Accountability In Shaping Behaviours of Humanitarian Actors

By Carlo Crudeli

This study analyses the role of accountability in shaping the behaviour of humanitarian actors. Accountability has become a key tenet in the sector and has been at the centre of the debate in recent years. However, the concept of humanitarian accountability in its current understanding is still relatively new, only emerging in the late 90s. While some efforts have been made to organize the existing knowledge on the subject by advancing new frameworks and reviewing existing literature on specific subtopics, this study represents the first review that systematically analyses the literature on the subject. This integrative review aims at identifying key concepts highlighted by the literature and provides a framework to better understand the idea of humanitarian accountability and how it can influence humanitarian actors.

The review confirms that accountability remains, today, badly defined in the humanitarian sector. Nevertheless, several concepts and ideas have been identified and explored, providing interesting pathways for action and reflection. After a brief summary of the history and evolution of the concept, the literature is reviewed to highlight the key questions and components of humanitarian accountability: who is it that should be held accountable? To whom should the latter be held accountable to? How is accountability implemented in practise and through which mechanisms and processes?

Finally, the study provides a synthesis of the current literature on the topic and proposes a certain interpretation and potential path of investigation, linking the discussion to other segments of contemporary literature such as public management, social psychology, and behavioural science. The review highlights the many contradictions and trade-offs which should be considered with regards to accountability. However, it also emphasizes the potential for significant effects on organizational and individual behaviours. Considering that both positive and negative effects have been identified, and the stakes of the humanitarian sector, a rigorous and honest reflection on the topic is not only advisable but required.

What Works? Corporate Volunteering by Design

By Sergio Pelaez

Most research on pro-sociality has attempted to find what’s behind our choice to help another people. Typical questions include: are we motivated by benevolence? Do we do it only if the cost is minimal, or the benefits great? Do economic incentives crowd out pro-sociality? Does it make us happier? Nonetheless, behavioural science is showing how our behaviour is not ruled by intentions rather than contextual cues.

This research aims to enrich the debate on pro-sociality, going beyond rational or moral reasons. Are we pro-socially hardwired or do we need to be nudged? In order to test it, I undertook a natural field experiment consisting of random e-mails inviting to make monthly donations to a Bank’s corporate volunteer programme. There are a wide variety of contextual elements that can be tested. For the purposes of this research, I have chosen to focus on a few of these, which I am grouping for practical purposes as forms of ‘social influence’: messenger, emotions and norms. The results obtained show how pro-sociality does not escape the thesis that ‘context matters’. The null hypothesis defined as the non-existence of differences between groups was rejected. It means that different contextual cues generate different outcomes in terms of interest and contribution to charitable giving. One of the key contributions of this experiment is that by splitting the treatment effect through different conversion rates (awareness, interest and donation), it allowed to evidence how a decision can change during the process if context is interrupted abruptly. These findings were categorized under two phenomena: ‘social proof failure’ and ‘social norms mixing with market norms’.

Hopefully, this research will contribute to the existing literature that seeks to understand what is behind our choice to help people, and show the contingent nature of pro-sociality.

Place Image Emergency Care: The Power of Logical De-Biasing and Food Culture Commonalities to Restore Post-Terror Place Image

By Rika Sakai

This research is aimed at finding effective interventions to reduce anticipated danger for visiting a place and also help people recover willingness to visit there, when the place image has been compromised by terror incidents.

The analysis is based on statistical significance found in a random control trial (RTC). This experiment is based on a survey commissioned from an international marketing research company and conducted on 1,000 Japanese speakers, within a month of the latest series of terrorist attacks in Europe in 2017. The survey asked the participants about their images and opinions of Europe.

People tend to judge a situation quickly without going through a rational thinking process by accessing the most easily available information or images; ‘availability heuristics/bias’ (Tversky & Kahneman, 1973). This bias is presumed to be present when people see major news outlet reports about security threats.

In order to eliminate such bias, providing easy statistical logic is presumed to be effective. Additionally, presenting stories of cultural interconnectedness is presumed to make people feel more curious and familiar with different places, consequently helping people be more willing to visit those places. This experiment is conducted to test these two hypotheses.

Three interventions are applied: availability heuristics/bias planting by showing web pages with a major newspaper’s headlines reporting recent terrorist attacks in Europe with a few photos; attempt to de-bias the availability heuristics by providing easy statistical logic; and/or providing a story of cultural interconnectedness of the world by using various similar dumplings.

The data demonstrated that all the intervention sets (de-biasing, food culture story, and the combination of both) were highly significantly effective to reduce anticipated danger. Especially the level of de-biasing availability heuristics/bias was statistically highly significant.

To help people recover willingness to visit, the best choice should be combining both of the interventions.

Fly-Tipping Behaviour and Social Norms

By Cian O Morain

Fly-tipping, which can be defined as the “illegal act of discarding waste in an unauthorised place” (Curran et al, 2007, p. 82), is an important public policy matter in the UK and elsewhere. Fly-tipped waste is costly to clean up (estimates from 2006-7 place the clearance cost in the UK at £73 million), causes environmental problems and promotes further anti-social behaviour. Likewise, it is an issue that is consistently highlighted by citizens as being important to them. Behavioural science has paid much attention to more casual littering behaviour, in particular the impact of social norms on subsequent behaviour. Littering behaviour is influenced by descriptive norms, which describe what most others do and are observed principally from the amount of litter in an area, and injunctive norms, which are the rules or beliefs that constitute the view of what one ought to be doing in a given situation. However, there has not to-date been a study investigating the impact of social norms on fly-tipping behaviour. This study outlines two proposed experiments which would do so. First, a survey is designed and administered to the area of Streatham South to identify existing injunctive and descriptive norms. Subsequently, these results are fed into two experiment designs aimed at influencing the decisions to leave the house with waste to be fly-tipped and to choose a specific location to fly-tip.



The Infidelity Temptation: Does Sexual Arousal Make Us More Prone to Cheat?

By Martina Höppner

The paradox between widely shared intentions of monogamy and high infidelity rates keeps fascinating both scientists and laypeople. Despite decades of research, this mystery remains unsolved. The infidelity literature has explored the relevance of relationship, demographic, and trait aspects, as well as contextual, situational, and environmental factors. The role that visceral factors in general and sexual arousal in particular may play in tempting individuals into infidelity has, however, hardly been explored. Using a randomised controlled experiment, this study investigates whether arousal-primed participants show more positive attitudes towards and higher intentions of infidelity compared to unaroused control group peers. To address potential social desirability effects, an implicit association test (IAT) is conducted. Several personality moderators are investigated in the exploratory part of the study. Arousal had no statistically significant impact on attitudes towards and intentions of infidelity for arousal-primed participants compared to control group peers. Nor did the IAT yield significant results.

This leaves two possibilities: Either, the arousal prime –despite inducing a statistically significant difference in arousal between treatment and control group participants –was not strong enough to tempt participants into changing their infidelity attitudes and intentions, thus supporting Loewenstein’s (1996) assumption that individuals can resist low levels of visceral factors effects. Or it is not that easy to induce individuals to change their infidelity attitudes and intentions, further supporting that monogamy is still the prevalent social norm. Of the explored personality moderators, only risk aversion in the health domain showed significant results with health risk averse individuals turning more positive towards infidelity when aroused, whereas health risk takers surprisingly turned more negative towards infidelity.

One of the major limitations of this study is that only infidelity attitudes and intentions could be investigated, not real-world incidents. Future studies should therefore explore how arousal –primed in different ways –influences actual infidelity.

Scheduled Spontaneity and its Effect on Wellbeing: Increasing the Sex Frequency of Couples Wanting More

By Enrique Gomez

Most people in long term relationships wants more sex than they currently have but are often prevented from pursuing more by their own behaviours.

This study used framing to change some of the behaviours stopping individuals in long term relationships (N=772) from actively trying to have more sex with their partners. The intervention modified a common perception where sexual problems are incorrectly attributed to exogenous causes; encouraging individuals to improve the sex lives in their relationships by scheduling sex rather than relying on spontaneity, whilst debunking common myths associated with scheduled sex. The intervention achieved significant treatment effects for sizeable sections of the treatment group: a 0.4 (t=2.59; P=0.01) time per month increase I the sex frequency of individuals wanting more sex and having an open mindset; going up to 0.8 (t=3.37, P=0.001), for men meeting the same criteria and even greater effects on those with higher levels of life satisfaction. The intervention also achieved a positive treatment effect by improving people’s willingness to talk to their partners about sex (t=2.47, P=0.014); evidencing increased intentionality and progress towards the implementation of scheduled sex.

This research contributes to the study of sexual, relationship and life satisfaction, offering new insights into the role that gender, happiness and openness can play in the quest of improved sexual satisfaction. The wellbeing benefits of scheduled spontaneity are not limited to sexuality and can extend to other fields of life. 


Social media

Risky News: Risk Perception Thrives News (and Fake News) Virality in Social Media

By Eduardo Garza

The days when the media controlled the flow of the news are long gone. Today, participants in social media platforms play a major role in making news ‘go viral’ by sharing, commenting or endorsing content regardless of its authenticity. This common practice is one reason why fake news travels faster than ever before. Fake news can influence our decisions and shape our beliefs in a wide array of topics such as war (Khaldarova & Pantti, 2016), discrimination (Banaji, Bhat, et al. 2019), terrorism (Ewart, 2012), climate change (Lewandowsky, et al., 2013) and vaccination (Chiou & Tucker, 2018).

Availability cascades are news ‘going viral’ as a result of a reciprocal information availability process between media and the public (Kuran & Sunstein, 1999). Social media currently generates these availability cascades. Hazardous activities that pose a risk, regardless of their probability of occurrence or scope of impact, are common topics of availability cascades. Thus, risk perception is expectedly one of the reasons underlying the public’s amplification of fake news in social media.

In this study, Paul Slovic’s categorisation of hazardous activities (Slovic, 1987) is divided into four risk perception types (RPTs), namely RPT-1 (‘not dread’, ‘known’); RPT-2 (‘not dread’, ‘unknown’); RPT-3 (‘dread’, ‘unknown’); and RPT-4 (‘dread’, ‘known’). The objective is to understand how RPTs, whether posed as true or fake news, have an effect on the propensity to share, comment or endorse news in social media. Understanding this phenomenon is of utmost importance for social media platforms, fact-checkers and policymakers to set the guidelines for reducing the spread of fake news in social media.

The study reveals several findings. First, share and endorse, driven by easiness, are users’ most preferred virality actions. Second, the effect of sharing and endorsing fake news is smaller than the effect of sharing and endorsing true news. Third, non-dread risks (RPT-1, RPT-2) are more likely to be shared than dread risks (RPT-3, RPT-4) probably due to perception shifts. Fourth, arousal, risk to oneself, risk to others and perceived accuracy are effective predictors of news virality. Finally, the results support the optimism bias and illusionary effect.

#Unethical Behaviour: Can a Tweet Make You Cheat Less? A Randomised Control Trial on Social Media and Dishonesty

By Nausica Rosina

Big corporates’ scandals such as Enron and Volkswagen make to the media, however cheating is part of everyday life of ordinary people as well. From cheating on taxes claimed to over-reporting the numbers of hours worked, people have many opportunities to lie. Although previous studies have proven that many behavioural interventions can reduce the level of cheating (Mazar N. et al., 2008; Shu L. et al., 2012; Ariely D., 2013; Zhang T. et al., 2014; Gunia B.C. et el., 2012; and many more), a dedicated research on the influence of social media has never been conducted. This study ran a between-groups randomised control trial (two treatments on social media vs no treatment) with 291 participants at the LSE Behavioural Research Lab. Contrary to expectations, social media did not decrease the level of cheating significantly, however descriptive statistical analysis suggests that social media could have an effect on reducing dishonesty.

The results of this study, although not significant, open the road to a new journey for behavioural science to tackle dishonesty with social media. Further studies with robust experimental framework are encouraged to promote effective and targeted interventions.



Hot Hands and Cold Feet: Perceptions of Goal-Scoring Probability in Premier League Football

By Simon March

The existence of a ‘hot hand effect’; the improved likelihood of a successful outcome following, andconditional upon, a previous successful outcome, has been a source of debate for over 30 years and it has been studied in numerous sporting contexts including basketball, baseball, golf and tennis. This study expands this discussion through the context of football (‘soccer’) and, using goal-scoring data from ten English Premier League seasons, it seeks to address the question of whether or not a hot hand effect exists among the elite players in the world’s most popular sport.

This scoring data is subsequently compared to survey data collected from 600 football fans with the goal of understanding whether belief in a hot hand effect exists among them and how this belief might evolveover the course of a goal-scoring streak. While this analysis finds no statistical evidence of a hot hand effect in football; scoring in one game does not appear to improve a player’s likelihood of scoring in their next game, the survey demonstrates a strong, contradictory belief among football fans that it does. A further survey of the same population finds that almost half of respondents also believe that not scoring in their previous game makes a player more likely to score in their next game, suggesting the presence of a ‘gambler’s fallacy’ counterpart to the hot hand belief when football fans judge the probability of a player scoring.

These findings are subsequently analysed in the context of existing hot hand and gambler’s fallacy literature and, in particular, the theory that both phenomena arise from a belief in the ‘law of small numbers’ and use of the associated ‘representativeness heuristic’, as first expressed in Gilovich, Vallone and Tversky’s seminal 1985 paper; ‘The Hot Hand in Basketball: On the Misperception of Random Sequences.’ While the evidence gathered in this study is compatible with Gilovich et al.’s broader hypothesis, these latest findings are problematic with respect to those authors’ assertion that it is the extent to which a sequence of results might be interpreted by a person as being random or non-random that causes these fallacious beliefs to occur. Several other existing explanatory hypotheses are discussed, and a new, alternative theory is proposed; that what is considered ‘representative’ in the context of the hot hand and gambler’s fallacies can be based on a person’s interpretation of contextual ‘reference points’, the influence of which may evolve, dynamically, over the course of a scoring/non-scoring streak. It is argued that, not only does this nascent reference-based model provide a valid behavioural explanation for the results of this study but they may, also, help to resolve certain inconsistencies in existing literature in this area. Furthermore, it is suggested, this model may offer a foundation for the long sought-after predictive framework that can determine which of the hot hand and gambler’s fallacies will be evoked in a given circumstance.




Negative Emotions Associated with Potential Betrayal of Trust Cause Us to Weigh Information Differently and Independently of Risk: Potential Economic Benefits from Increased Access to Algorithm-Driven Digital Services

By Britt Spyrou

This research demonstrates that access to algorithm-driven digital services increases economic wellbeing because it offers us an environment which instinctively appeals to our desire to avoid negative emotion. 

Using a vignette developed by Cubitt et al (2017) a representative sample of 584 smartphone users in the United States were asked to choose a ride from the airport: a more expensive safe fixed-price option or a cheaper risky meter-based option which could be more expensive one in five times. The decision-maker about an identical risk was alternated between the control (weather) and two treatments (human and algorithm). Participants chose the safe option most often and more quickly when presented with a human decision-maker (choosing the safe option 14% more often and half a second quicker than the control). When participants were presented with an algorithmic decision-maker they chose the safe option 10% more often and 0.3 seconds more quickly than the control, costing them less money than the human condition. The impact of the distrust signal in the vignette on economic well-being was then tested through two-way dictator and trust games. Nearly all results from the games were not different from chance. However, a significant difference was found for passengers in the algorithm condition who earned 2.3 tokens more in the trust game compared to the passengers in the human condition. This was due to the general absence of negative emotion in the algorithm condition.

This finding is based on an implicit association test which compared reaction times to associate the concepts of 'trust' and 'distrust' with 'human' and 'algorithm'. This test found that the participants naturally associated 'algorithm' with 'distrust', finding it took 15.5 milliseconds to associate the concepts. In contrast it took participants 176.5 milliseconds to associate 'human' with 'distrust'. The difficulty in making this association manifested itself in participants in the vignette for the human condition making an economically irrational decision to avoid experiencing that negative emotion. The aversion to negative emotion was even more pronounced for the subgroup where both parties made a profit in the games. These participants chose the safe option 17% more often in the human condition and 12% more often in the algorithm condition than in the control condition. This research confirmed that negative emotions associated with potential betrayal of trust cause us to weigh information differently and independently of risk. One implication of this research is that higher sensitivity to negative emotion could be associated with greater potential for economic prosperity.

Alternatively, these findings also suggest we could lower the intuitive reactions and increase economic wellbeing by presenting people with options which invoke less negative emotion, such as access to algorithm-driven digital services. Overall, the findings support further research on the utility of digital inclusion initiatives to boost economic development in high distrust contexts like pandemics, conflict and corruption.



Reflections on Workplace Reflecting

By Shelley Hoppe

Modern office workers face multiple distractions in a polychronic world that encourages reactive over reflective behaviours. Research has shown that an increasingly pervasive use of technology is leading to less deliberate thinking, impacting both workplace behaviours and workers’ well-being. This study aims to investigate whether reflection can be prompted just before a task to help increase the time spent reflecting on that task, as well as whether or not this additional reflection will improve performance. Building on existing work within the science of interruptions, it asks: Will visual and mindset priming result in increased reflection time, and in improved performance on a Cognitive Reflection Test? In this context, the study aims to test several interventions that seek to actively prompt employees to increase reflection in relation to a specific task.

Using the COM-B Model (Michie et al., 2011) as a framework, and based on are view of the literature on priming, both visual and using self-efficacy-mindset enhancing questions, an online experiment was performed with 602 employed adults based across the UK. Respondents were randomly divided into three conditions (visual priming, self-efficacy mindset priming and a combination of both) and asked to complete a Cognitive Reflection Test. Analysis of the time spent and the overall scores on the test demonstrated that it was possible to increase participants’ reflection time, but that this did not significantly improve their test scores. The results indicate that the treatments may have impacted some individuals more than others, but it is unclear why. On this basis, the results suggest that it is possible to motivate some people to spend more time reflecting by using a combination of visual and selfefficacy mindset primes. Further research is needed to identify whether or not additional reflection is beneficial and also whether or not if affects different personality types in different ways.

Improving Transfer of Learning: The Use of MCII to Enhance Leadership Development

By Andreas Priestland

In the UK, organisations spend a significant amount of time and money on training their staff. But the equation of ‘training + staff = improved performance’ does not always appear to add up. There is a gap between the inputs and the outputs. It seems that what people learn often does not get transferred into what they do afterwards.

This paper looks at what might be some of the causes of the gap and what might be done to help the employee bridge it. Following their attendance at a one day leadership workshop, this study involved employees of an international pharmaceutical company being exposed to mental contrasting with implementation intentions (MCII) as a way to assist them focus on a clear goal and initiate action for post-workshop implementation. After a period of two weeks individuals were contacted for a second time in order to test for differences in implementation, time to implementation, and outcome success.

No significant differences between the control and experimental groups were detected. Substantial attrition and non-response rates lead to a loss of sample size and low statistical power. Additional potential contributory factors are discussed, along with avenues for further exploration.

Which Frameworks Work at Work? Understanding the Efficacy of the Most Widely Used Behavioural Science Frameworks for Non-Academic Practitioners

By George Smith

Several behaviour change intervention frameworks now exist to aid non-academic practitioners in applying the techniques from academic behavioural science literature in their work without the need for academic study. However there has to date not been a scientific study into the efficacy of these frameworks in the hands of non-academic practitioners – who make up a large proportion of their users as behavioural science’s popularity has increased. This study aims to fill this gap in the literature by laying out three measures of framework efficacy: rigour of guidance, comprehension and recall of the framework, and the quality of the frameworks’ non-academic users’ intervention designs. It assessed the frameworks against the latter two measures in a randomised control trial. After a pre-survey which uncovered that the top three most-used behavioural science frameworks were MINDSPACE, EAST, and COM-B, 190 working professionals were randomly assigned to study one of these 3 frameworks, then invited to create interventions for 2 behavioural challenges using them. After, they were also asked to recall the outline content of framework they used. Their interventions were judged by a behavioural scientist for quality against the APEASE criteria and quantified for analysis.

The study uncovered that while there was no significant difference between the quality of the interventions non-academics produced using the three frameworks, EAST and COM-B were significantly easier for participants to recall. However, since these frameworks do not sufficiently advise their users’ of the latest criteria for what a behaviour change intervention is said to require, the study suggests a new behaviour change intervention framework which deliberately covers these criteria, that is also easy to comprehend and recall could be the most effective at allowing non-academic practitioners to achieve significant behavioural results, and with it, facilitate widespread positive behaviour change.

Behavioural Change in Organizations: Can Salience Improve the Effectiveness of Institutional Recognition Programmes?

By Juan Carlos Valverde

This research explores the effects of recency in improving the effectiveness of an existing recognition programme at a private institution. To create a recency effect, randomly selected staff members within 211 software development teams were exposed to seven internal communications with a reference to the recognition programme, whilst a control group was exposed to the same communications without such device. The experiment led to an 8.3% (p=0.048) marginally significant increase in the adoption of the behaviour promoted by the recognition programme. The effect is reduced to (8.04% p=0.05) after controlling for the covariates identified. The experiment supports the hypothesis that recency could be effective in improving the impact of recognition programmes, thus contributing to accelerate behavioural change in institutional contexts.

Psychological Safety and Learning Behaviour in Virtual Work Teams

By Rafael Ortega

Over the last few years, a growing body of research has examined psychological safety – a shared belief about how safe it is to take interpersonal risks - and its consequences for individuals, teams, and organisations. However, the role of psychological safety in virtual teams may diverge from the one it plays in teams interacting in face-to-face settings. In addition, most studies have been conducted in English-speaking countries and in Western cultures characterized by low levels of uncertainty avoidance. By adjusting Edmondson´s model of team learning behaviour (Edmondson, 1999), I explored antecedents and outcomes of team psychological safety in virtual work teams in a school in Colombia, a Spanish-speaking country with a national culture characterized by high levels of uncertainty avoidance. Results for 30 virtual working teams suggest that psychological safety and its antecedents of team leader coaching and team leader inclusiveness are of relevance in this context. However, in addition to psychological safety, there could be other factors acting as mediators between its antecedents (i.e., team context support, team leader coaching, and team leader inclusiveness) and the outcome of team learning behaviour. Finally, psychological safety may not lead to team learning behaviour when cultural norms and performance expectations don´t require it.


If you have any questions about the Executive MSc Behavioural Science programme, please contact pbs.emsc@lse.ac.uk. Alternatively you can find answers to the most common questions about the programme on our FAQ page.