The effects of gender and ethnicity on absolute vs. relative ratings for low-fat school lunch entrées

School lunch interventions often have to compete with less healthy alternatives. This study was designed to test the absolute and relative preferences for three low-fat entrées currently available within three schools in central Texas. Students rated three low-fat school entrées: chef salad, turkey sandwich and baked potato. Two rating formats were used: (1) absolute ratings, scored individually, with five-point scales; and (2) relative ratings indicated by a series of dichotomous, 'yes-no' forced choice questions. Across three elementary schools, 238 fifth grade students received parental consent to participate. One hundred thirty-three (56%) were present for all taste tests. The final subject pool was comprised of 77.4% Hispanic and 22.5% Caucasian children. The absolute and relative ratings were analyzed via 2 (gender)×2 (ethnicity)×3 (entrée) repeated measures analyses of variance. Relationships among these variables were assessed with bivariate correlations. Absolute ratings were positive, averaging 4.05 on the five-point scale, with no differences as a result of ethnicity or gender. Relative preferences were lower, indicating they may be a better measure of intention to select. A significant effect for food [ F (2,126) = 4.09, P = 0.01], as well as for food×gender interaction, [F (2,126) = 8.03, P < 0.01]was found. Post hocs revealed lower ratings for chef salad compared with the other low-fat entrées. Girls reported greater intention to select salad than boys (d=0.62). These data suggest that although low-fat items can be acceptable to children, absolute ratings may be a poor indicator of behaviour when children are provided a choice among competing entrées.