Revisiting the Self-Refilling Bowl of Soup

Source: PippiLongstocking/iStock

Quite a bit of research in psychology has been devoted to understanding the factors that influence eating behavior. This work is important because taking in calories at an amount that balances calories burned is an important part of the process of maintaining a consistent weight.

A complicating factor in eating behavior is that when people choose to start and stop eating is based both on internal factors (like a feeling of hunger) as well as external factors (it is around dinner time). As a result, it is important to understand the contributions of these internal and external factors.

One prominent study that aimed to demonstrate the role of external factors was done by Brian Wansink and his colleagues in 2005. In this study, the researchers compared the eating behavior of two groups of participants. One group ate soup from a bowl, and periodically, research assistants would refill that bowl when it was nearly empty. This group had a visual cue that they had eaten most of their bowl of soup. A second group ate from a bowl that automatically refilled as the soup was eaten so that participants did not get any visual cue about how much they had eaten. This study found that participants who ate from the self-refilling bowl ate more soup than those who ate from a bowl that was drained as they ate and occasionally refilled by a research assistant. Interestingly, the study found that there was no difference between groups in how much they thought they had eaten or how hungry they were after eating.

This result is fascinating and (unsurprisingly) got a lot of coverage in the popular press. It was also discussed in many books and popular accounts of how to eat in a healthy way.

Unfortunately, investigations of Wansink’s work led to a number of findings of misconduct and sloppy research practices, which ultimately led to the retraction of a number of his published scientific papers. In this context, there has been skepticism about whether this intriguing finding is real. Indeed, the authors point out several places in which the original manuscript is unclear in ways that lead to some concerns about the original findings.

A 2024 paper in the Journal of Experimental Psychology: General by Alejandra Lopez, Alyssa Choi, Nadia Dellawar, Brooke Cullen, Sonia Avila Contreras, Daniel Rosenfeld, and Janet Tomiyama did a pre-registered replication of the initial study with a larger sample of participants: 464 completed the study successfully.

Happily, the main findings of the original paper held up in the replication, though the primary effect was smaller than what was observed initially. As before, participants who had the self-refilling bowl ate more than those who could see their bowl emptying. Interestingly, the participants who could see their bowl emptying actually predicted that they ate more soup than those who had the self-refilling bowl. The groups did not differ significantly in how hungry they were after eating.

This replication suggests that the main finding of the original paper can continue to influence how we think about eating behavior. Factors like an empty plate influence when people think that they are done eating. This research suggests that when people are trying to control the amount they eat, putting a portion on a plate or bowl and then finishing that portion can help them know when they are done eating. These external cues may often be more accurate than a feeling of hunger—particularly because it can take a while before food that is ingested makes a person feel like they have eaten enough.

This study also demonstrates the importance of replication. It is unfortunate when there are cases of scientific misconduct that lead the community to call findings into question. It is particularly unfortunate when findings that turn out to be reliable (and could have strong practical benefits) are discounted because of the reputation of the researcher who did them. A replication by an independent team can help the field determine which results merit continued attention.