Recognizing the need for an objective, reliable and economic method for assessing behaviors indicative of maladjustment, Pimm and McClure (1966) empirically constructed a checklist of 100 observable behaviors deriving from teacher descriptions of maladaptive classroom behavior. Items of the checklist were then related to age, sex, I&, and to clinical assessments of behavior disorder. This paper reports a further analysis of the checklist in a search for homogeneous clusters of items which might be said to represent basic dimensions of problem behavior. PROCEDURE SubjectsSs for the original stud were 1,445 school children in the Ottawa Public Schools. These children were complete samples of a i the grade one classes in 23 schools in one geographic half of the city. As problem behaviors were infrequent for the sample as a whole, Ss for the current study were those 484 males and 343 females for whom a minimum of three items on the checklist had been marked as being present by their teachers. This reduction in Ss was a necessary first step in reducing the item pool to those items present in at least 10% of the cases. MethodFrequency counts were obtained for the 827 remaining Ss on each of the 100 items of the checklist; 36 items met the 10% frequency criterion. Intercorrelations of the items were then obtained and subjected to a principal axis factor analysis with communalities estimated by use of the highest correlation in the column. Factorization was terminated when the eigenvalues fell below unity; a varimax rotation was subsequently performed on the factors extracted. RESULTSRotated factor loadings are presented in Table 1 for those 31 of the 36 variables which had at least one loading of .25 or greater. The four factors extracted accounted for 71% of the communality.Factor one appears to represent a cluster of behaviors that would be considered troublesome by first-grade teachers, but would be characteristic of the classroom behavior of nursery school or even many kindergarten children. None of the behavioral items appears particularly pathological and we have interpreted this factor as reflecting simply behavioral immaturity. The second factor loads primarily a group of items involving verbal behavior with smaller loadings on three variables involving motor activity. This factor may well represent poor control of impulses to activity, but in view of the larger loadings for the verbal items we have labeled it verbal overactivity. Factor three clearly involves behavior contrary to most social and moral standards and seems a clear representation of the frequently found (Peterson, 1961;Quay & Quay, 1965; Quay, Morse, & Cutler, 1966) conduct problem dimension. Although the loadings are small, Factor four seems obviously to reflect the anxious-withdrawal or personality problem constellation also found in the prior analyses cited above.

120 5"s at each of 6 developmental levels performed an instrumental lever-pulling task. 6 reward groups (0%, 10%, 30%, 50%, 70%, 100%) were formed within each age level and measures of response speed were taken. The typical finding of faster response speeds for partially as compared with continuously rewarded groups was not evident across all ages. Furthermore, some partial reward schedules other than 50% also lead to more vigorous responding as compared with 100% reward. In terms of frustration theory, it was suggested that desire to win and problem-solving strategies employed at different age levels, as well as expectancy for reward, may all play important roles in determining reaction to nonreward.