As psychologist Benjamin Scheibehenne and his wife left the restaurant where they had just finished dinner, they discussed whether to stop somewhere else for dessert. It was an everyday decision, one they had made countless times before, but this particular evening they could not make up their minds.

"When we came out of the restaurant, we didn't really know whether we were still hungry or not," Scheibehenne recalls. "We realized we were completely clueless about how much we actually consumed."

The couple's appetite for dessert owed its ambivalence to the unusual nature of their dining experience: The Scheibehennes had visited a "dark restaurant," where sight-impaired waiters serve customers their meals in a total blackout—a trend that claims to enhance the sensory experience of eating, and which has gained popularity in Europe and Asia, with some inroads into the U.S.

Scheibehenne, a psychologist at the University of Basel in Switzerland, realized that dark restaurants could provide a great setting for an experiment about how visual cues influence the way people estimate portion size and evaluate hunger. The eventual outcome was a new study published online August 13 in the journal Appetite, which suggests that an accurate judgment of satiety depends more on what we see with our eyes rather than what we put in our stomachs.

"The main result is that it seems surprisingly difficult to estimate the amount of food you consume in the absence of visual information," Scheibehenne says. "The feedback your body gives you is not working very reliably. It certainly works when you overeat—when you're stuffed—but in most situations, people need visual cues to estimate quantities and hunger."

David Zald, a psychologist who studies hunger at Vanderbilt University, was impressed with the new study. "This idea that we use visual cues to influence our eating rather than it being a strict feedback from the stomach is interesting," he says. "There have been some previous studies that hinted at this, but I think it's a good extension of their findings and it's a particularly elegant demonstration."

Along with colleagues Peter Todd of Indiana University Bloomington and Brian Wansink of Cornell University, Scheibehenne invited 64 participants over the course of two days to a free lunch in a dark restaurant in downtown Berlin. Sight-impaired waiters guided the participants through complete darkness to their tables and asked the diners to surrender any cell phones, luminous watches or other sources of light. After the participants dined on two main courses (vegetable risotto and goulash with noodles) in the dark, the waiters returned their guests to the lighted portion of the restaurant and offered dessert—a platter of fruit slices and cheese on toothpicks from which the diners could serve themselves. Unbeknownst to the participants, the researchers had provided one group of diners with normal-size lunch portions for the blind meal, whereas the other group received supersize versions.

The experimenters weighed each plate before it left the kitchen and after it returned from the diners' table to assess how much the participants ate. They also counted the number of toothpicks discarded to measure dessert consumption. Finally, the researchers asked the participants to estimate how much food they had consumed and to rate their satiety. Scheibehenne and his colleagues then repeated their study with 32 different participants and one major change: These volunteers ate their meals in a well-lit restaurant—they could see everything.

The study found that estimating how much you eat in the dark is much harder than doing so in the light, as anyone might expect. But darkness also changed how the participants judged their own feelings of hunger and how much dessert they ate. For those who dined in the light the amount of dessert consumed correlated with their lunch sizes: Participants who had regular-size portions ate an average of 12 fruit sticks, whereas those who had supersize portions ate an average of eight fruit sticks. But in the dark, portion size lost its relevance. Those diners who had consumed supersize portions in the dark ate an average of seven fruit sticks, whereas those who consumed regular portions ate an average of eight fruit sticks—hardly a large difference.

"What this shows is that the capacity for how much we eat is driven by what we see, not necessarily by how much food is actually there," Wansink says. "Your stomach is a very imprecise instrument when it comes to measuring food—you might be able to tell roughly whether it's full or not, but you can't tell if it's full of meat or butter, calories or lettuce."

Scheibehenne adds that vision has the advantage of immediacy, whereas our stomachs need more time to evaluate what we digest. Both Scheibehenne and Wansink also explain that over the course of our lives we build up a reliable visual memory that links different portion sizes to different levels of satiety. "We can think, 'Hey, I am usually full when I eat a full plate of food,'" Wansink says. "From 40 years of past experience, I know if I eat this I will probably be full—that's a much more reliable assessment than our stomachs give us."

"I think it's a really clever design," Zald says. "They were able to simulate an eating situation quite well, and they measured quite a few different variables." The new study's design, however, presents two main potential confounding factors. First, the act of eating in the dark may have fundamentally changed the typical dining experience in more ways than the researchers intended. Most participants who dined blind reported at least some difficulty locating their food and using silverware, and—compared with diners who ate in the light—said that they paid much more attention to the taste of the unidentified food than to its weight or portion size.

Second, participants who ate lunch in the well-lit restaurant were likely influenced by social cues that diners in the dark did not have. Participants with supersize portions in the light, for example, may have felt somewhat self-conscious about eating more than participants who had smaller portions, and may have subsequently resisted taking more fruit sticks than their peers. Scheibehenne says that he and his colleagues tried to control for this unintended peer pressure by having participants eat dessert with a different group of participants than they lunched with.

Susan Roberts, a nutrition researcher at Tufts University School of Medicine, says she does not question the findings but thinks the unusual nature of the experiment should not be overlooked. "People were given food not really of their choice in complete darkness—it's just not relevant to normal eating at night," Roberts says. Moreover, the study group itself, in which 74 of the 96 participants were university students, may not be a good sample, she suggests. "College students and free food are not a great combination."

Nevertheless, Scheibehenne feels the study offers general lessons. "It's important to pay attention to things in the environment that could trick you into thinking a particular portion size is smaller or larger than it really is, or distort your visual cues," he points out. Consider how eating in front of the television keeps our eyes on a screen instead of on our plates, for instance.

Wansink echoes this concern: "When your food is out of sight, it's out of mind. And this doesn't just happen in a dark restaurant. If you serve yourself small portions deliberately, you could be in tremendous danger of repeatedly refilling your plate and overeating."