Eating vegetarian is not like voting

Heavy-handed social pressure works better if there actually is a widespread social norm encouraging something.

A canonical finding in political science: if you tell people 1) whether they vote or not is public information and 2) you plan to tell their neighbors whether they vote or not in an upcoming election, you can increase voter turnout quite a bit — from about 29.7% at baseline to 37.8% for a 2006 primary in Michigan.

Dannenberg, Klatt & Weingärtner (2024) study the effects of social norms and observability on food choice. The experiment invited German university students to take part in a survey where they “could win a voucher for three sandwiches at a local restaurant as a thank-you gift and asked to choose either vegan, vegetarian, meat, or no voucher.” The setup looked like this:

In one treatment condition, first-year students were informed that, in a group of students already enrolled, a large percentage reported that they followed a vegan or vegetarian diet. In another treatment condition, they were informed that only a small minority reported to follow a vegan or vegetarian diet. No information about dietary habits was provided in the control treatment. As a second randomized treatment variable, we manipulated observability by informing half of the participants that they would have to publicly announce their voucher choice if they won, while the choice of the other half of participants remained private.

If eating vegetarian/vegan was amenable to social norms/pressure the way voting is, you’d probably expect that the combination of learning that 1) “a large percentage [of your fellow students] reported that they followed a vegan or vegetarian diet” and 2) everyone would then see what you’re eating would result in a lot more non-meat meals.

Alas, no dice:

Figure 1, Dannenberg, Klatt & Weingärtner (2024)

Apparently the problem is men, who “show little tendency to follow the social norm regardless of whether they are confronted with a high or low proportion of vegetarians and vegans, or whether they are observed or not. When faced with a high proportion of vegetarians and vegans and observed, they even show a slight tendency to choose more meat.” Women in the sample were more amenable to social pressure/publicity, but not enough to produce a meaningful overall effect.

It matters how plausible the social norm is

My theory for the divergence here is that you can’t expect to convince a random person in a western society that being vegetarian/vegan is normal. It’s not. It’s actually tricky to pin down how many Germans are vegetarian, but a meta-analysis on this question suggests something in the 1-4% range.¹ Let’s say that 3X as many college students are vegetarian as the upper bound of that, so 12%. That’s still very uncommon. The high-veg treatment groups in the study were told that “60% of students in that group reported eating a vegetarian or vegan diet.” I am pretty sure I would not believe you if you told me that. Even if I believed it about a particular group, I would still know how common meat consumption is in the world at large. A group with 60% vegetarians would be downright weird. Would I want to conform to that group? Per this study, it seems most people don’t.

By contrast, about 60% of eligible voters in America typically vote in presidential elections. Even if they don’t bother to vote in a midterm, ~90% of Americans say voting is a very or somewhat important part of “being a good member of society.” Imagine telling people that you eat meat sometimes, and then imagine telling them that you didn’t vote in the last election. Which do you think will get people mad at you?²

There’s a shared understanding that voting is a normal, good thing. Norms around vegetarianism are much weaker.³ Claiming that a vegetarian norm exists, when people have a lifetime of experience telling them otherwise, is likely to be a hard sell.

1. The actual range in the studies is 0.96% to 11.2%, but a lot of the surveys that get estimates on the higher range are worded to include people who “predominantly” adhere to a vegetarian diet — I don’t know what that means — and also pescatarians.

2. I speak from experience: in 2008 I didn’t vote because I was in Scotland and I didn’t feel like it, and boy did people not like that! Most societies have norms against telling people off to their faces — hence the need for gossip — but in my milieu, saying “I don’t vote” creates an immediate imperative to tell the offender that they’re bad, ideally with great gusto.

3. To be fair — 🎵to be fair🎵 — there’s some evidence that surveys on vegetarianism overreport its incidence, which suggests that at least some people think it’s a good, desirable thing to be even if they’re not quite there yet. I’m totally willing to buy that there are some scenes where being a vegetarian is cool, and some where it’s not. (I also recognize that this essay is a bit loose with the distinction between norm-as-in-normal (descriptive) and norm-as-in-desirable (injunctive).)