Imagine that you're buying a box of chocolates for a loved one on Valentine's Day, and the store is offering a free gift if you buy a slightly inferior box of chocolates instead of the best chocolates in the store. Which do you buy?

You are more likely to go for the inferior box, according to new research from the University of Chicago Booth School of Business, published in the Journal of Consumer Research.

In the study, "The Friendly Taking Effect: How Interpersonal Closeness Leads to Seemingly Selfish Yet Jointly Maximizing Choice," by Chicago Booth Professor Ayelet Fishbach, University of Florida's Yanping Tu, and University of Chicago's Alex Shaw, the researchers find that people are more likely to take from a close other than a distant other. In a series of studies, the researchers determine that this tendency is rooted in a friendly intention of trying to maximize the total benefits for the pair, or the so-called "self-other collective."

"We dub this the 'friendly taking effect' because such overt taking behavior is actually rooted in a friendly intention, that is, concern for the total benefit," the study says.

Just as people are more likely to give more to close friends than to strangers, people may be more likely to give less to close friends than to strangers if there is a mutual overall benefit for doing so.

Call it altruistic selfishness. The gift-giver may see himself and a close friend as a unit--and choose the best total gift for the unit rather than for either individual. The total gift could end up including the gift purchased, a free gift, and any money saved.

The researchers conducted seven experiments that explored different kinds of relationships--including naturally occurring friendships and those developed in the lab. The activities in the experiments included sampling chocolates, getting massages, sharing cabs and choosing airline mileage programs.

The results of each experiment support the hypothesis that, because people focus on total benefits when making decisions about how to allocate resources between themselves and people they're close to, they choose the option that benefits themselves.

In one experiment, 63 University of Chicago undergraduates were invited to enter a raffle in which each winner and one person they knew would sample gourmet chocolate truffles. They had to choose between two differently distributed prizes: in Package A, the winner would receive seven truffles and the other person three, for a total of 10 truffles; in Package B, the winner would receive two truffles and the other person six, for a total of 8 truffles. The truffles could not be shared.

When participants knew they'd sample the truffles with someone they felt close to, almost two-thirds chose the package where they would receive more truffles. But when participants anticipated sampling truffles with someone they felt less close to, the fraction flipped, and only about one-third chose that option.

The findings have implications for merchants, retailers and marketers. The research suggests that rewards and other bundled incentives could be particularly influential to people making purchases for close acquaintances.

Although consumers generally spend more on gifts for people they're close to, the researchers say, "they might also be more influenced by discounts, sales, and other saving opportunities" when buying for those loved ones.