快猫短视频

Who you choose to eat with influences what you decide to eat, according to a study of 38 million food purchases at a university campus over eight years.

Kristina Gligori膰 at the Swiss Federal Institute of Technology in Lausanne and her colleagues tracked purchases made on campus using smart cards. The team tracked the eating habits of 39,000 anonymised students and staff from 2010 to 2018. On average, people鈥檚 purchases were tracked for 578 days and they visited shops, cafes, restaurants and vending machines 188 times.

Of the 39,000 users, 830 people were identified as 鈥渕atched pairs鈥� of strangers after their purchases were monitored for a year and found to be similar. Their food purchases then deviated as they respectively began to eat with different friends.

鈥淚n order to really isolate the effect of others on our choices, we need millions of purchases and tens of thousands of individuals to be able to find really comparable people,鈥� says Gligori膰. She thinks of the people in matched pairs as 鈥渄oppelgangers whose history is the same, but one by chance happens to start eating with someone who is a healthy eating partner, and someone starts eating with an unhealthy eating partner鈥�.

Their purchases were then monitored to see how they changed before and after they began eating socially, measured by calculating when and where purchases were made, and ensuring they were close enough to be in the same queue as others. The researchers discovered that people pick similar items to the person they are eating with. If a new friend eats pizza, the tracked person is more likely to eat pizza.

In comparison to the other person in their matched pair, people whose eating companions ate more unhealthy foods were more likely to buy one extra soft drink and 0.5 additional pizzas in the six months after they buddy up. The same happened with healthier food: people who made friends with healthy eaters bought on average an extra 5.71 healthy items and 1.13 fewer unhealthy options over six months.

A food鈥檚 healthiness was measured by a nutritionist, who ranked items purchased based on Swiss dietary guidelines. The calorie content of the foods purchased wasn鈥檛 calculated because some items were registered on receipts as something as vague as 鈥渓unch meal鈥�, which can be made up of two or three options of a main course with a carbohydrate, a salad and water.

The study is an important addition to nutritional research, says Zeinab Mulla at Imperial College London. While plenty of research exists about social influences on diet, it is often self-reported data 鈥� where people frequently claim to be healthier than they are 鈥� or based on small laboratory exercises. 鈥淒igital tracing is where we鈥檙e trying to take the field of dietary assessment,鈥� she says.

However, the paper didn鈥檛 account for one key element of the student diet: the packed lunch. 鈥淲e don鈥檛 know what people cook at home and bring to campus,鈥� says Gligori膰, who hopes to incorporate that in future studies.