John Menzies, Jan Michael Bauer, Lucia Reisch
The Nudge-it “Policy meets Research” Workshop took place on 18th May. Organised by Prof Lucia Reisch and her team at Copenhagen Business School, the aim of the meeting was to discuss in a diverse forum some evidence-based insights from the Nudge-it project.
The morning panel discussion (“Thinking for policies: Gathering research evidence”) was opened by Prof Michele Belot and Prof Todd Hare. They focused on the importance of developing tools to understand how food choices are made. Acknowledging that currently the mechanisms behind food choice are not well-understood, they went on to present potential anchors for interventions based on evidence from data collected in the Nudge-it project.
This was followed by open debate. Limits on food reformulation, questions around whether findings can be generalised to all socio-economic groups, and the importance of aligning the diverse interests of scientists, policy-makers and food manufacturers were discussed. The concept of “Epicurean” nudging - nudging towards pleasure and enjoyment of food (as opposed to arguably more “puritanical” nudges towards healthier behaviour) proved to be a repeating theme of the entire workshop.
Rene van Bavel (Joint Research Center, European Commission) discussed applying outcomes from behavioural sciences to EU policy. He insisted that an evidence-based and nuanced understanding is needed to develop effective policy, and we must acknowledge the power of bias and heuristics, particularly using the insights from Daniel Kahneman’s ideas https://en.wikipedia.org/wiki/Thinking,_Fast_and_Slow about how humans make decisions – often using “system 1” (fast, intuitive and easy) rather than “system 2” (slow, analytical and effortful). In most cases, Rene argued that policies need to be developed under the guidance of our understanding of system 1: more information is not necessarily better. Rene also highlighted that the link between evidence and policy is often indirect, and needs a mass of evidence. A one-to-one relationship between individual findings and policy development ought to be avoided.
Marieke Adriaanse (NUDGIS and WINK Projects, University of Utrecht) provided insightful and productive criticism of the intervention anchors proposed earlier in the meeting, indicating the importance of taking advantage of habitual behaviour. Importantly, habit means we avoid having to exert self-control. This may sound negative, but if the habits are beneficial and self-control is difficult, the outcome of habitual behaviour can be positive. Habit-forming nudges could prove to be a means of maximising the power of behavioural interventions. An interesting point was raised on calorie information labelling. While overall evidence on the effects of calorie consumption is mixed, some studies show that labelling can also lead to increased calorie buying. For instance, in food-insecure communities, often linked to low socio-economic status, people may be aiming to obtain as many calories per $ spent.
Lunch involved a mini-replication of a study looking at the usual default of opting in to vegetarian food (which is often healthier and more sustainable produced compared to meat), and comparing this with opting in to non-vegetarian food – in other words, a nudge where vegetarian food is the default. In the original study, only 2% of the participants opted in to eat vegetarian food, but 87% chose vegetarian food when it was the default. In our workshop, only 3 participants chose non-vegetarian food with no apparent deficit in enjoyment.
The afternoon panel discussion (“Thinking about policy: How to make use of the evidence?”) opened with Knut Karevold and Samira Lekhal (GreeNudge). Small daily increases in energy intake can have large effects on body weight over time, and Samira discussed the reverse – showing how small interventions can make big differences. GreeNudge uses evidence-based tools to help people make healthier and sustainable choices. Samira discussed a number of initiatives including competence training for health care professionals that resulted in an increase in the number of health professionals following national guidelines for treating obesity in children. Knut discussed the promotion of understanding of behavioural approaches to non-experts, and discussed results on a meta-analysis of behavioural insights on food choice.
Romain Cadario (IESEG Paris, France) discussed a meta-analysis of field experiments, showing that behavioural nudges (exploiting system 1 thinking: making healthier options more convenient and using a smaller plate size, are most effective. Interestingly, Romain’s data also indicated that it is easier to reduce unhealthy eating than increase healthy eating. To strengthen the evidence-based of behavioural interventions, Romain and his co-author created a website [link to: https://insead.eu.qualtrics.com/jfe/form/SV_1ENmOXnDHtoYnkx] to enter with new study-results to their data base and thereby keep their meta-analysis up-to-date.
Mario Mazzocchi (University of Bologna; Editor of “Food Policy”) focused on evaluating the effectiveness of policy, arguing that strong evidence to guide policy is still hard to come by. He discussed particular problems with the time-frames of evaluation and identifying causality between effects, rather than a simple association. An intriguing proposal was to conspicuously label taxed foods (the so-called “sin taxes” like a sugar tax) and non-taxed foods in supermarkets, display them in separate aisles (the “taxed aisle and the “non-taxed” aisles) and observe customers’ choice behaviours.
Pelle Hansen (The European Nudge Network and iNudgeyou, and a self-described “guerrilla scientist”) gave an overview of some unpublished field experiments in nudging food choice behaviour. Of particular interest was data indicating that obesity researchers are markedly worse at estimating the calorie content of foods compared to the general public.
The highlight of the workshop was a keynote by Prof Cass Sunstein (the Robert Walmsley University Professor at Harvard Law School, and alongside Prof Richard Thaler – a recent recipient of the Nobel Memorial Prize in Economics – the originator of nudging). Cass spoke about nudges that fail. Again, this may sound negative, but we learn from our mistakes, and understanding what contributes to the failure of interventions allows us to design better interventions.
So-called “counternudges” can cause nudges to fail. Often these are measures taken by private businesses to overcome changes driven by behavioural interventions. For example, mandating fast food restaurants to provide small side-dishes by default as part of combo menus could be easily “counternudged” by waiters prompting customers to supersize their meals.
Strong antecedent preference can cause nudges to fail - nudging doesn’t affect strong preferences. If an individual knows what she wants to eat, nudges that reduce or increase the prominence of a food in a supermarket or restaurant don’t have an effect on choice. However, this can be considered on utilitarian grounds: by failing to thwart this individual’s enjoyment of her food, and assuming eating the food increases to some extent to her overall happiness, this failed nudge has not reduced social welfare and can be thought of as a “fortunate fail”.
Directly related to this, nudges fail if they are welfare-reducing. This can be difficult to determine, especially when peering into the future, guessing at unintended consequences of proposed interventions. Cass gave an overview of discussions in the US around whether foods sold in cinemas should be calorie-labelled in the same way as foods in supermarkets. Eventually this was implemented and despite Cass not being part of the final decision, a policy of reminding popcorn eaters of the unhealthiness of their chosen snack was enough for one of his colleagues to accuse him (jokingly) of “ruining popcorn”. An illustrative example highlighting potential welfare-reducing effect of behavioural interventions beyond the intended effects on food choice.
Cass suggested that social welfare should be our guide to developing behavioural interventions, while remembering that short-term losses (not eating a cake) may lead to long-term gains (not becoming overweight). We can also use “altering rules” – techniques to ensure deliberative judgement. These can fortify decisions provoked by nudges. We can also use framing rules to make counternudges less powerful, or use countercounternudges. However, we need to stay aware of the intrinsic value of choice, the power of reactance [link to https://en.wikipedia.org/wiki/Reactance_(psychology)], avoiding using confusing information in nudges, and compensating behaviour – especially in health nudges where we may choose to eat healthily in the morning but “compensate” by eating a donut in the afternoon.
In summary, the participants represented a great mix of stakeholders involved in policy-making and food-decision research. Reaching from neuro-scientists to public administrators the workshop sparked interesting discussions and reflections about linking scientific evidence about the determinants of food choice with the requirement for and practical application of real-life food policy.