Ultra-processed foods (UPFs) have become public enemy number one in nutrition debates. From dementia to obesity and an epidemic of "food addiction", these factory-made products, including crisps, ready meals, fizzy drinks and packaged snacks, are blamed for a wide range of modern health problems. Some experts argue that they're "specifically formulated and aggressively marketed to maximise consumption and corporate profits", hijacking our brain's reward systems to make us eat beyond our needs.
Authors
- Graham Finlayson
Professor of Psychobiology, University of Leeds
- James Stubbs
Professor in Appetite & Energy Balance, Faculty of Medicine and Health School of Psychology, University of Leeds
Policymakers have proposed bold interventions: warning labels, marketing restrictions, taxes, even outright bans near schools. But how much of this urgency is based on solid evidence?
My colleagues and I wanted to step back and ask: what actually makes people like a food? And what drives them to overeat - not just enjoy it, but keep eating after hunger has passed? We studied more than 3,000 UK adults and their responses to over 400 everyday foods. What we found challenges the simplistic UPF narrative and offers a more nuanced way forward.
Two ideas often get blurred in nutrition discourse: liking a food and hedonic overeating (eating for pleasure rather than hunger). Liking is about taste. Hedonic overeating is about continuing to eat because the food feels good. They're related, but not identical. Many people like porridge but rarely binge on it. Chocolate, biscuits and ice cream, on the other hand, top both lists.
We conducted three large online studies where participants rated photos of unbranded food portions for how much they liked them and how likely they were to overeat them. The foods were recognisable items from a typical UK shopping basket: jacket potatoes, apples, noodles, cottage pie, custard creams - more than 400 in total.
We then compared these responses with three things: the foods' nutritional content (fat, sugar, fibre, energy density), their classification as ultra-processed by the widely used Nova system - a food classification method that groups foods by the extent and purpose of their processing - and how people perceived them (sweet, fatty, processed, healthy and so on).
Perception power
Some findings were expected: people liked foods they ate often, and calorie-dense foods were more likely to lead to overeating.
But the more surprising insight came from the role of beliefs and perceptions. Nutrient content mattered - people rated high-fat, high-carb foods as more enjoyable, and low-fibre, high-calorie foods as more "bingeable". But what people believed about the food also mattered, a lot.
Perceiving a food as sweet, fatty or highly processed increased the likelihood of overeating, regardless of its actual nutritional content. Foods believed to be bitter or high in fibre had the opposite effect.
In one survey , we could predict 78% of the variation in people's likelihood of overeating by combining nutrient data (41%) with beliefs about the food and its sensory qualities (another 38%).
In short: how we think about food affects how we eat it, just as much as what's actually in it.
This brings us to ultra-processed foods. Despite the intense scrutiny, classifying a food as "ultra-processed" added very little to our predictive models.
Once we accounted for nutrient content and food perceptions, the Nova classification explained less than 2% of the variation in liking and just 4% in overeating.
That's not to say all UPFs are harmless . Many are high in calories, low in fibre and easy to overconsume. But the UPF label is a blunt instrument. It lumps together sugary soft drinks with fortified cereals, protein bars with vegan meat alternatives.
Some of these products may be less healthy , but others can be helpful - especially for older adults with low appetites, people on restricted diets or those seeking convenient nutrition.
The message that all UPFs are bad oversimplifies the issue . People don't eat based on food labels alone. They eat based on how a food tastes, how it makes them feel and how it fits with their health, social or emotional goals.
Relying on UPF labels to shape policy could backfire. Warning labels might steer people away from foods that are actually beneficial , like wholegrain cereals, or create confusion about what's genuinely unhealthy.
Instead, we recommend a more informed, personalised approach:
• Boost food literacy: help people understand what makes food satisfying, what drives cravings, and how to recognise their personal cues for overeating.
• Reformulate with intention: design food products that are enjoyable and filling, rather than relying on bland "diet" options or ultra-palatable snacks.
• Address eating motivations: people eat for many reasons beyond hunger - for comfort, connection and pleasure. Supporting alternative habits while maximising enjoyment could reduce dependence on low-quality foods.
It's not just about processing
Some UPFs do deserve concern . They're calorie dense, aggressively marketed and often sold in oversized portions. But they're not a smoking gun.
Labelling entire categories of food as bad based purely on their processing misses the complexity of eating behaviour. What drives us to eat and overeat is complicated but not beyond understanding. We now have the data and models to unpack those motivations and support people in building healthier, more satisfying diets.
Ultimately, the nutritional and sensory characteristics of food - and how we perceive them - matter more than whether something came out of a packet. If we want to encourage better eating habits, it's time to stop demonising food groups and start focusing on the psychology behind our choices.
Graham Finlayson has received funding from Horizon Europe, UKRI and Slimming World, UK.
James Stubbs consults to Slimming World UK. He receives funding from UKRI.
