This study compares and contrasts teachers' and parents' construal of children's behaviour, drawn from the general population rather than clinical extremes. Questionnaire data were obtained from parents and teachers on the same 3-7 year old children via nurseries and mainstream schools. Item reduction followed standard psychometric approaches. Varimax-rotated principal components analysis was undertaken on dichotomised teacher and parent responses, giving a four-factor solution as most appropriate both for teachers and parents (aggression; anxiety; context-directed behaviour; social maturity/ communicativeness). The low correlation found between parents and teachers on corresponding factors suggests strong divergence of perspectives and/or situation-dependence of behaviours. Reliability was high enough to not provide the explanation - i.e., serial correlations over three months within each respondent type were relatively strong. However, slight changes also took place over time: teacher ratings of anxiety differentiated out into anxiety proper and social maturity, while ratings of aggression and non-context-directed ('inappropriate') behaviour integrated towards a single factor which may reflect threat to classroom order. Implications for parent-teacher communication are discussed.