Acceptabilly of school-served food items was assessed by three methods: a five-point hedonic scale (HED); a five-point scale on which students estimated the amount they had eaten (AMT); and weighed plate waste. Reliability of HED scale was highlv sienificant: reliability of AMT scale was significant for all but one item. AMT scores. had closer relationship to food consumption than did HED scores. A statistical model was constructed for predicting food consumption from ratings; it was possible to obtain an indication of consumption by elementary students from either scale. Application of methods is discussed. Use of student ratings for predicting consumption by elementary students seemed feasible in test-retest situations.