SciTech

Robot helps reveal science behind snack choices

Jodi Forlizzi, a professor in the departments of design, human computer interaction, and computer science, poses with “Snackbot,” a snack-delivering robot aimed to help understand how people choose between different snacks. (credit: Photo courtesy of Jodi Forlizzi) Jodi Forlizzi, a professor in the departments of design, human computer interaction, and computer science, poses with “Snackbot,” a snack-delivering robot aimed to help understand how people choose between different snacks. (credit: Photo courtesy of Jodi Forlizzi) Credit: Josh Smith/Staff Credit: Josh Smith/Staff

Snacking in the workplace is a common ritual, and many employees may be concerned with making the right choice when it comes to their waistlines and overall health. A group of Carnegie Mellon researchers are trying to determine the best way to encourage people to make healthier snack choices.

Researchers in the Human-Computer Interaction Institute recently developed a system that helps people make self-beneficial snacking decisions in the workplace based on theories of behavioral economics, which differs from traditional economics by assuming that people do not make the most rational decisions all the time. The research was recently published in Proceedings of the 2011 Annual Conference on Human Factors in Computing Systems.

“Obesity is a growing problem worldwide,” said Jodi Forlizzi, one of the study’s authors and a professor in the departments of design, human-computer interaction, and computer science. “In our university labs and office buildings, people work long hours, often substituting snacks for meals. Having the option to choose a healthy snack is a small step towards ameliorating these problems.”

In a study involving 61 participants, the researchers first examined participants’ tendency to favor a “default option” in which a snack choice was made for them, and they could receive an alternative option only if they specifically asked. These results were compared to participants’ tendencies to rationally weigh pros and cons of a snack choice if they were asked to plan ahead.

For two consecutive days, experimenters delivered apples or cookies to participants’ offices in the afternoon. On the first day, everyone was asked to choose between apples and cookies. One group, labeled the “default group,” was told they would have apples the next day unless they specified they wanted cookies, and a “planning group” was asked to choose a snack they would like for the next day.

About half of the participants in the default and planning groups chose an apple on the first day, but this percentage increased on the second day as nearly two-thirds chose apples. More participants made the switch in the planning group than the default group, implying that the planning strategy may have had a larger effect. Researchers also asked the participants to rate their snacks and found that both groups indicated less enjoyment than a control group that was presented with the same snack choices as the first day. Possibly, the researchers thought, people felt an implicit pressure to have a healthy snack or simply didn’t like planning ahead.

In a second experiment, the researchers observed the effects of using a robot to deliver snacks to the participants. In one delivery method, apples and cookies were evenly laid out on a tray, making both choices equally accessible to participants. In another method, apples were delivered on a tray while the cookies were placed in a bag underneath the tray, making the cookies slightly less convenient to participants.

The researchers found that making the apples slightly more accessible was effective in influencing people to make the healthier choice. Furthermore, people’s enjoyment level was generally unaffected by this robot delivery method versus being asked if they wanted a different snack than the default option. The researchers reasoned that this might be due to the lack of social pressure in the robot experiment.

A last experiment revealed that information such as calorie counts presented to participants had little effect on their snack choices, and in many cases, the information caused participants with a healthy diet to choose unhealthy snacks. The researchers proposed the possibility that since a candy bar has fewer calories than participants may have thought, they allowed themselves to indulge in junk food.

The research is just one illustration of how simple changes in user interfaces can have significant impact on people’s choices. It implies that there is great potential for behavioral economics to benefit human health.