REFERENCE: Clayton, B., Hovell, M. F. (1985). The validity of a role-played measure of student's snack selections. Education and Treatment of Children, 8, 229-238.
Two measures of nutritious snack selection were compared in order to judge the validity of a role-played measure of snack selection. Elementary school students were provided a free-time snack period before and after nutrition education instruction. For half of these snack periods, students selected and ate snacks. For the remaining periods, food models (pictures) were provided, and students were asked to select those snacks they would like to eat, if real. A Pearson Product Moment correlation coefficient was calculated for snack and picture selections. None of these correlations was greater than .48, and most did not reach statistical significance. It was concluded that the model snack procedure was not a valid measure of nutritious snack selec-tion. It was suggested that explicit tests of validity may be important for other role-played and analogue measures of behavior.