Climate swaps made online carts healthier. Health swaps didn't clearly cut carbon.

Randomised controlled trial (RCT). 1,201 people. 3 online shopping visits about 1 week apart (May-June 2025): 1 baseline visit without intervention, then 2 visits in the assigned arm.

A trial. Chance decided who did what.

Who: US adults 18+ recruited via Prolific (convenience sample, ~50% aged 18-25 by quota; mean age 31.5; 53% women; more Black, more college-educated and lower-income than the US population)

• Health swaps: health+4.7 ptsscore out of 100
• Climate swaps: carbon-39%lower footprint
• Climate swaps: health+2.4 ptshealthier too

More to know, from our read of the paper:

• Simulated online store: participants did not pay for or receive their groceries (25 were promised groceries, later paid $50 instead). Effects in real or brick-and-mortar stores unknown
• Outcome is the healthfulness score (Ofcom, 1-100) and carbon footprint of selected items, not what people ate or any health outcome. Only 2 intervention visits about a week apart
• Labels (A-F grades) and swaps were always delivered together, so their separate effects cannot be told apart
• Carbon reductions (~39% climate swaps, ~33% combined) are the authors' approximate percentages vs a baseline mean of 8.5 kg CO2-eq/kg
• All 3 swap types increased negative emotions (guilt, worry) while shopping

In a randomized trial, 1,201 US adults shopped three times, a week apart, in a realistic test online grocery store. Some were offered swaps to healthier foods, lower-carbon foods, both, or none. Health swaps raised the health score of their picks by 4.7 points out of 100. Climate swaps cut the carbon footprint by about 39% and also made picks healthier (+2.4 points), but health swaps alone did not clearly cut carbon. People accepted only about 1 in 4 swaps, and nobody paid for real groceries, so real stores may differ.

We read the abstract and the full text of this paper and checked every number on this page against it (17 facts, checked on 2026-10-29). The study design is named above. Cohort and other observational studies are described as linked, not proven. Found an error? Email theis@onestudyaday.com.