To determine how well the relative preference ratings of foods by groups of soldiers correspond with ratings by groups of civilians under similar test situations, the ratings for twelve test foods made by four 50-man groups of soldiers were compared with the ratings made by two 40-man groups of civilians. Under the conditions of the experiment, correlations above .9 were found between the soldier and civilian ratings, and it was concluded that "...pretesting of rations in the Institute laboratory may be considered equivalent to pretesting by the soldier -consumer panel method.".