# What the evidence says about menus

One page for anyone who has to justify a menu change. Every claim carries its strength rating, its limitations and its sources, including the trials that found nothing.

*The Balance of Health. Generated 2026-09-06. Sample data pending verification.*

## Increasing the share of plant-based options on a menu increases the share of plant-based meals sold, without reducing total sales.

Yes, in canteen settings. Doubling vegetarian availability raised vegetarian sales by roughly 8 to 15 percentage points, and almost nobody noticed the change.

Evidence strength: moderate. Two field studies in different countries agree on the direction and the mechanism, but both are university catering, the two estimates in the original study differ by roughly a factor of two, and at least one menu-nudge trial in a similar setting found nothing at all.

In a study of 94,644 meals across university cafeterias, raising the share of vegetarian options from a quarter to a half raised vegetarian sales by about 15 percentage points observationally and 7.8 points in the experimental arm.

A French replication found the odds of a vegetarian choice rose about 2.6 times, with slightly higher satisfaction, and only about 6% of diners noticed anything had changed.

The mechanism is not persuasion. It is that people choose from what is in front of them, and most of them are not making a considered decision at all.

- Limitations:
- Both studies are university catering with a captive audience. A supermarket, a pub or a wedding is a different problem.
- Nykänen and colleagues found that dish-of-the-day and menu-sequence nudges changed neither dish choice nor the weight of meat served.
- Plan with the experimental figure, not the observational one.

## Naming a dish for its flavour rather than for what it lacks or for its health properties increases how often it is chosen, with the food unchanged.

Yes. Identically prepared vegetables were chosen 41% more often under a flavour label than a health-restrictive one, and 'meat-free' performed worse than almost every alternative tested.

Evidence strength: moderate. Replicated across dining halls and commercial partners with a plausible mechanism, but mostly single-setting trials measuring selection at the counter, and the industry-partnered work is less fully reported than peer-reviewed studies.

Stanford dining-hall trials served the same vegetables under different labels. Flavour-focused labels beat health-restrictive ones by 41% and basic ones by about 25%.

The Better Buying Lab found 'meat-free' underperformed nearly every alternative name. Describing a dish by what is missing invites loss aversion and makes the taste harder to imagine.

The same effect runs the other way: people rated an identical mango lassi as markedly less enjoyable when it was labelled healthy.

- Limitations:
- Measures what people select, not what they eat or what they do next time.
- University dining halls and partner restaurants are specific settings.
- Naming is not free: a name that oversells the dish will cost more in disappointment than it gains at the counter.

## Sources

| Source | Type | Quality |
| --- | --- | --- |
| Impact of increasing vegetarian availability on meal selection and sales in cafeterias — University of Cambridge (2019) | randomised-controlled-trial | strong |
| Increasing vegetarian availability in a French university restaurant — French university catering research group (2023) | randomised-controlled-trial | moderate |
| Nudging towards plant-based main dishes in a university restaurant — University research group (2022) | randomised-controlled-trial | moderate |
| Increasing vegetable intake by emphasizing tasty and enjoyable attributes — Stanford University (2019) | randomised-controlled-trial | moderate |
| It's not easy eating green: the language of sustainable diets — World Resources Institute, Better Buying Lab (2020) | randomised-controlled-trial | moderate |