TEACHERS' RATINGS are now widely accepted as a valid procedure for the appraisal of children's behavior and personality. Many of the problems which have traditionally complicated the use of rat ings , e. g., confusions in trait meanings, halo ef fects, and the lack of reference standards, have been recognized and resolved adequately. Other problems, however, such as the components of the ratings, sex and age differences in subjects' behav ior, and the external criteria for validity have not been studied adequately. In the analysis of teacher ratings it should be recognized that it is the teachers' ratingbehavior which is being analyzed. To the extent that their rat ings are an accurate reflection of the child's traits then validity can be attributed to the rating instru ment. This may be a logical conclusion since the rating instrument structures their responses in a manner which is meaningful for them, and which is also rooted in the child's behavior. Cattell and Drendahl (6), however, report seri ous deficiencies in teachers' ratings. They found that some students actually possessed characteris tics which were in direct contrast to those rated by the teacher. Research by Barron (1) indicated that ratings are an inefficient, and perhaps inadequate procedure for the identification of certain personal ity traits. In spite of these problems Cattell (4), Cattell and Coan (5), Digman (7) and others have re lied heavily on teachers' ratings in the formulation of personality theory. Several studies have examined the validity of teachers' ratings for specific traits (12, 2 , 10). Their findings tend to validate the use of teacher rat ings for certain types of in-school conduct, and ad justment behavior. These studies sugge st that teacher ratings have specific rather than general validities. For example, Leton (10) found that the highest correspondence between teachers' ratings and scores on a personality inventory, existed for scales which measured achievement motivation, in tellectual efficiency, sociability, and responsibility. In similar studies Holland (8), Ryan (11), and Tal lent (13) found a strong association between teach ers' ratings and students' achievement. They re ported such traits as leadership potential, self-con trol, and responsibility as holding high associations among students' ratings. The Winnetka Scale for Rating School Behavior is perhaps the only such instrument for which behav ioral incidents in classroom situations were used as preliminary criteria and for which factor validi ties were established in the original standardization (16). The pilot testing of the scale was carried out on 250 children ranging from nursery school through sixth grade. The revised form was administered to 1200 children in elementary schools in Winnetka, Chicago, and rural Kansas. The first analysis of the scale extracted three factors which were iden tified as Pupil Desirability, Leadership and Social Consciousness (14). The third factor was regarded as a bipolar factor with Self-Sufficiency or Social Independence as the opposite aspect. The first fac tor was a general factor in which the factor lo ad ings ranged from . 36 to . 76 for the 33 scales. This factor was subsequently identified as Cooperation. In the following year Thurstone (14) presented an improved procedure for factor extraction and the analysis of the Winnetka data was repeated. Four factors were then identified with the addition of a