There are times I wish I had a macro for the beginning of a post on obesity. Some way to say obesity is bad, obesity is prevalent, and nothing seems to work. You know the drill.
But there’s a new study in Health Affairs that was surprisingly promising:
We performed three related field experiments at a single fast-food restaurant to determine whether these reported sentiments could be translated into a strategy to alter calorie consumption. All of the experiments addressed three important elements of eating behavior.
First, do people spontaneously request smaller portions—that is, even if smaller portions are not specifically noted as an option on a menu or signage? Second, do people accept explicit spoken offers to take smaller portions in order to reduce calories? Third, does taking a smaller portion of one meal component lead to indulgence in other meal components, so that the calorie “savings” from downsizing are immediately lost?
Each experiment addressed an additional question. In experiment 1, we explored whether offering a nominal (twenty-five-cent) discount for downsizing would result in more customers’ accepting the offer than offering no discount. In experiment 2, we examined whether offering an opportunity to accept a smaller portion would be more effective than providing calorie labels in encouraging moderation. In experiment 3, we investigated whether downsizing appealed only to customers who would otherwise have thrown away uneaten food, thereby affecting calories ordered but not calories consumed.
Let’s start with experiment 1. First, they measured how many customers would spontaneously request a smaller portion of a high-calorie, high-starch side dish. Not surprisingly, only 1% did. But if customers were asked, on the other hand, one third accepted the offer, regardless of whether a discount was offered. What’s more, those that did downsize did not compensate by up-sizing any other portions of the meal. Those that downsized ordered significantly fewer calories, 100 fewer on average.
Continue reading…
