EATING BEHAVIORS OF RESTRAINED EATERS, OVERWEIGHT AND NORMAL ADULTS.

A study of human eating behavior in laboratory and natural environments was undertaken to answer three questions: (1) Do manifestly obese subjects (Manifestly Obese, MO), non-obese restrained eaters (Latent Obese, LO), and non-obese non-restrained eaters (Normal, NM) differ in specific short-term intake behaviors as measured in the laboratory? (2) Do the three groups differ in self-reported 14-day habitual energy intake patterns? (3) Are there systematic variations of energy consumption within and outside the laboratory in subject groups? Twenty two adult female subjects were assigned to three groups and studied in both settings. The purpose of the laboratory study was to investigate the effects of energy content and sweet taste on subsequent consumption behaviors of a variety of palatable foods. All subjects thought they were participating in a taste panel. The protocols were designed to minimize cognitive cues on amount eaten. Laboratory results showed that all subjects adjusted subsequent energy intake according to the energy content of a preload given half an hour before. However, the groups responded to the sweet taste in the preload differently. The LO consistently lowered energy intake of testmeal after a sweet preload. They overestimated their own consumption most of the time, and deviated even further with sweet taste stimulus. The MO and NM groups did not exhibit these responses. Results of self-reported 14-day dietary records showed that the LO had the highest daily variation of energy intake, the most intake from evening meals, and the greatest within group variation among the three subject groups. Comparison of mean and variation of energy intake within and outside the laboratory suggested that the observed consumption of the LO subjects in the laboratory was the least representative of the habitual intake, while that of the NM group was a satisfactory measure. These systematic group differences in two settings could be due to differential degrees of externality, measurement errors, or reactivity to measurement procedures