We tend to choose healthier food for ourselves than we do for others, a study shows.

RODALE NEWS, EMMAUS, PA—This just in from the checkout line: If you're shopping for yourself, it's more likely to be healthy. But if you're shopping for yourself and others besides, take a quick look in your basket before you get to the checkout line—there may well be an extra box of cookies or container of ice cream in there. That’s right, a new study published in the Journal of Consumer Research reveals that shopping for other people may be interfering with your ability to make healthy food choices.

THE DETAILS: Study author Juliano Laran, PhD, assistant professor of marketing at the University of Miami, had a group of university students take a series of four computerized tests. The tests asked the students to select foods for themselves, for others, and for an anonymous "Mr. A." The result was that the students were always more likely to make healthier food choices for themselves, selecting items like raisins, celery sticks, and Cheerios, whereas they’d make more indulgent, less-healthy choices for others, buying such items as chocolate bars, cookies, Doritos, ice cream, and doughnuts.

In the case of the anonymous Mr. A, Laran asked the students specifically about buying either a granola bar or a chocolate truffle. They picked the granola bar for themselves nearly 60 percent of the time, and Mr. A received the granola bar only 40 percent of the time.

In a related field test, Laran interviewed members of the public as they exited a grocery store. He analyzed their receipts, and also asked who they were buying food for. Sixty-five percent were buying food for other people, while 35 percent were buying only for themselves. As with the student tests, people buying for others had more indulgent, unhealthy items in their baskets than people shopping for themselves. And it didn't matter who the recipients were. Those shopping for their families were just as likely to include indulgent food as those shopping for friends or roommates.