Consistency of food intake over four eating sessions in the laboratory

The aim of this study was to test the consistency (repeatability) of laboratory measures of food intake, including eating microstructure (cumulative food intake curves), in non-obese humans across four eating sessions, each separated by 1 week. A secondary aim was to test the effect of dietary restraint on the food intake of female participants. Mixed model analyses were used to compare average food intake across sessions, and intraclass correlation coefficients (ICC) were employed to compare within subject variability to total variance. High and low restraint females and low restraint males consumed four lunches in a Universal Eating Monitor (UEM) laboratory. The lunches consisted of one type of sandwich (chicken salad sandwich squares) or three types of sandwiches (chicken salad, ham, and turkey sandwich squares) presented in counterbalanced order. Measures of food intake were stable for men and women, regardless of sandwich variety. In females, level of dietary restraint (high vs. low) did not significantly affect food intake. The results indicate that eating behavior in the laboratory is sufficiently stable over time to justify evaluation of interventions designed to alter food intake using within subject designs.