Learning Lessons: How to Practice Nudging around the World

"There may be good reasons to use psychological techniques to help improve behavior from increasing our saving for retirement to eating more healthily, but without a systematic means of reporting the psychological methods used by governments across the world, researchers, practitioners and the public will continue to remain in the dark about how they are implemented, how effective they are, and what benefits there are to the individual and the state." - Magda Osman

Many governments across the world use psychological and behavioural economic techniques (nudges) to steer us towards making beneficial decisions. A nudge is an example of a behavioural change intervention (BCI). Typically, a particular problem (e.g., public littering), in a particular context (e.g. parks) is identified by a practitioner (e.g., policymaker), and a specific behaviour is targeted (e.g., increasing the use of litter bins in parks) through a BCI (e.g., paint the litter bins in attractive colours to motivate people to use them more often). This article aims to highlight the importance of accurately cataloguing the types of BCIs that have been trialled/implemented across the world by: (i) introducing the Organisation for Economic Co-operation and Development (OECD)'s 2017 report, Behavioural Insights and Public Policy Lessons from Around the World, (ii) presenting the details of a simple classification system that was applied to all 111 case studies of BCIs referred to in the OECD report, and (iii) discussing what needs to be done to help advance practitioners' pursuit of effective BCIs to resolve social policy issues.

As reported here, the academic community has focused on answering questions around the theoretical foundations of nudge-type BCIs. BCIs based on behavioural insights, as the OECD understands them in its 2017 report, "are an essentially inductive approach to policy making that uses psychology (however it is referred to by the OECD) to empirically determine how people make decisions (rationally or otherwise): the OECD treats behavioural interventions as 'policy initiatives which are designed explicitly on previously existing behavioural evidence and/or based on a new experiment'...."

The OECD report presents the 111 cases along the following criteria: (i) country the BCI took place in, (ii) policy domain, (iii) start and end date of the BCI, (iv) objective of the BCI, (v) type of BCI used, and (vi) application of the BCI. As a support to practitioners and researchers, Section 2 of the article presents a summary of the core details extracted from the 111 cases included in the OECD report. A link to the full summary table is provided, which classifies the 111 cases along 19 criteria that were developed by the authors of this article.

In short, the analysis demonstrates that none of the techniques used has scientifically proven effectiveness. Of the 111 cases, 16% did not qualify as psychological and behavioural economic methods, and 20% were designed for the purposes of gathering information about behaviour associated with a particular policy issue, not for investigating how to change behaviour. Fifty-eight percent of the 111 were not found to have the anticipated effect on behaviour, and only 18% were later implemented by governments. No details about the cost of the initiatives were given, and where the interventions were positive, there is no information on the scale of the impact. Therefore, according to the authors of the article, these omissions mean that policymakers are unable to make any practical judgement of cost effectiveness.

They recognise that the OECD report is "a laudable attempt to provide basic details about the use of BCIs worldwide. By doing so, it sets a precedent for other organisations to do the same, which in the long run will help advance practitioners' and scientists' understanding of BCIs and their effectiveness. However, the report is substantially limited because efficacy cannot be determined in any meaningful way from the details included in the report." In Section 3 of the article, they lay out the main reasons for why effectiveness cannot be determined, and why this is problematic. For instance, "The most common details that appear in the OECD report regarding statistics are percentage changes in behaviour after the implementation of the BCI....Simply reporting the percentage change of the implementation of a BCI on behaviour means next to nothing without knowing the inferential statistics that were used to compare against the baseline/control condition (if it was even included) or another intervention, and the effect size of the change."

The authors suggest that "Researchers and practitioners would benefit from adopting an agreed classification system. For instance, clinical medicine has the Template for Intervention Description and Replication (TIDieR), for better reporting of interventions....This check list and guide is supposed to standardise reporting and aid replicability. We advocate going further than this, by also standardising the reporting of results, in order to be able to assess BCIs and their effectiveness, and details about cost should also be included for the sake of complete transparency."