Many people seek to reduce their meat consumption for a variety of reasons, including concern for animal welfare, the environment, or their health. Policymakers looking to encourage this behavior frequently turn to “nudge” policies. These are interventions that aim to shape people’s choices without directly constraining them, mandating one particular choice, or using direct penalties or rewards (such as monetary incentives) to drive behavior.

In recent years, the flurry of interest in nudges in the 2000s and 2010s has come under significant scrutiny. Better statistical methods, larger, longer studies, and attempts to combat structural issues like publication bias (where positive findings are more likely to be published) have undermined many initially promising results across policy areas. However, the world of nudges is large and varied. Some nudges likely do work in some cases; it’s just a case of carefully sifting through the evidence to find out what works, and where.

Food choice is one area that’s been singled out as particularly promising. Thus, this meta-analysis was carried out to dive into the details and find out which food choice nudges work, for whom, and under what conditions. The study focused on efforts to reduce meat consumption in foodservice settings such as canteens, restaurants, and cafeterias.

The researchers carried out a search of published literature, identifying 78 nudges reported across 33 studies in 28 papers. They used an established framework to classify the nudges into three broad categories, which were further divided into nine specific techniques.

The team then used a technique called moderator analysis to test what effect different factors had on the nudge’s effectiveness at reducing meat consumption. These factors included the type of nudge, demographics of the study participants, and study design. They also tested a number of ideas about how the analysis was likely to pan out. For instance, they expected changing choice defaults would be the most effective nudge, and that nudges would be most effective for younger people, women, and in study samples with fewer veg*ns. They also anticipated that nudges implemented over longer periods of time would show less effect as people tend to revert to their previous behaviors, and that menu-based nudges would be more effective than off-menu strategies like posters or staff recommendations.

As predicted, changing choice defaults turned out to be the most effective nudge, reducing meat consumption by 54% on average. Indeed, this was the only nudge that worked at all. Three nudge types, including changing option-related effort, changing option consequences, and providing reminders, simply had no evidence available. The remaining five types had no statistically significant impact on meat consumption (in fact, a few very slightly increased it, but again, not to the level of significance).

Furthermore, no notable differences could be discerned based on age, gender, or share of veg*ns in the study sample. Length of intervention did make a small difference, but in the opposite direction to what was expected, with nudges implemented over a longer period of time showing larger reductions in meat consumption. Menu-based nudges performed slightly better than off-menu ones as predicted, and studies that used a control group showed somewhat better results than those that simply compared the same participants before and after the nudge.

However, it’s crucial to note the study’s limitations. Just 12 of the 78 nudges were of the one type that worked. This means that the researchers’ attempts to study the impact of the moderating factors were effectively performed with a sample size of 12, diluted with 66 nudges not showing any effect for reasons unrelated to those factors. As a result, the small effects found for these analyses must be taken with caution.

It’s also worth noting that the study didn’t report any correction for multiple comparisons — a standard concern when testing many moderators simultaneously. It could be that the small effects observed are purely a coincidence driven by how many of the 12 nudges had those factors.

Despite the study’s limitations, the key conclusion is strong: changing choice defaults works well, while none of the other nudge types were found to be effective. Therefore, efforts to reduce meat consumption in foodservice settings should emphasize this tactic. The researchers recommend a graduated, multi-step policy approach:

1. Mandate at least one veg*n option per meal in public canteens;
2. Make that option the default;
3. Give public recognition to foodservice settings that have successfully implemented the default nudge; and
4. Provide training for foodservice managers on how to implement default nudges.

While further research is required to understand the nuances of changing choice defaults — such as how best to deliver this nudge and who it works best for — the policy takeaway for now is straightforward and obvious.