Food at checkouts: support varies with the message, not proof of effectiveness

An experiment on how a food policy is presented

The policies studied set nutrition standards for products offered at store checkouts. According to the scientific abstract, they have been introduced in three US jurisdictions. The experiment does not test their effects in those stores: it compares the arguments used to present them.

Participants were randomly assigned to one of four messages. These emphasized expanding healthy choices, reducing parental stress by removing unhealthy products from children's view, children's and community health, or tackling predatory marketing.

Recruitment used an online convenience sample, with quotas intended to approximate national sociodemographic characteristics. This method does not guarantee a sample representative of all US adults.

Majority support, but sensitive to the argument

Overall, 65% of participants supported the policy. Support reached 68% with the parental stress message, compared with 64% for the predatory marketing message and 61% for the message about expanding healthy choices. The abstract reports statistically significant differences for these comparisons.

The message focused on children's and community health received 67% support, also higher than the message about expanding healthy choices. The results presented do not establish that any one message outperforms all the others.

Support does not tell us what people will buy

Random assignment makes it possible to study the effect of message framing on self-reported support in this experiment. It does not demonstrate that the policy would change customers' purchases, food consumption or health.

The distinction matters: this study tells us about a measure's acceptability, not its nutritional effectiveness. Nor does it allow the observed levels of support to be directly extrapolated to French consumers.

What we can take away

Takeaway: benefits for parents and children appear to encourage support for checkout nutrition standards. To know whether these standards improve diets, we need findings on purchases and consumption, which are absent from the dossier.