Memory, Sex, and Food: The Price of Error

False memories aren't just problematic in court. They cause trouble elsewhere in society, and particularly in medicine.

As Elizabeth Loftus became known for her work on memory errors, survey researchers sought her help in order to understand how such errors were skewing responses to questionnaires. In crime surveys, memory tended to exaggerate local crime rates, which, in turn, could distort the allocation of law enforcement. In political surveys, people falsely recalled having voted in previous elections, thereby polluting samples of "voters." The biggest area of concern was health behavior. People didn't accurately remember which diseases they'd had, which precautions they'd taken, or how often they'd brushed their teeth. As a result, they misled their doctors and skewed public health data.

Sex was a good example. Loftus and her colleagues found that people exaggerated their consistency in using condoms and underreported their sex partners. Consequently, such people underestimated their degree of risk. Loftus worried that this tendency would corrupt the personal medical histories doctors used to diagnose and prescribe treatment for patients. It would also mislead intervention programs by falsely minimizing problem behaviors.

Rosy memories plagued food research, too. A typical survey, Loftus and her coauthors noted, might ask, "How often have you eaten chicken in the last 12 months?" Again, self-flattering answers—people reporting that they had eaten more spinach or less chocolate than was really the case—could skew medical histories, distort the prevalence of risk factors, and lead public-health programs astray. They could also contaminate science. "Suppose, for example, that you are a researcher interested in the relationship between fat in the diet and the development of breast cancer," Loftus wrote. "You decide to interview a group of women who have developed breast cancer and a group that is cancer free, asking questions about their diets and eating habits." Some of these women "might exaggerate the amount of healthy foods they ingested and minimize the quantity of unhealthy, fatty foods that also made up part of their diet." As a result, "you would be led to the wrong conclusions."  

And these were just memory's natural errors. By inducing additional errors, therapists could further distort public data. Suppose they persuaded their clients that the clients had been bullied as children. Such false memories, Loftus and her colleagues observed, might inflate estimates of the prevalence of bullying.

What if these two problems converged? What would happen if therapeutic distortion ventured into the realm of food memories?

Despite her own warnings, Loftus would soon lead the way.