Summary.This paper reports a series of studies in the development of a Study Behaviour Questionnaire (SBQ), for use with university students. The major problem in the present research was to determine optimal ways of combining SBQ item scores: by sorting people's responses to items (R‐analysis), or by sorting items on the basis of how they discriminate between people (Q‐analysis).464 Arts and Science first‐year students completed the SBQ. The responses were analysed in both R and Q designs, and the resulting scale scores for each student were merged with his high school and university performance data. In the R‐analysis, 102 items were selected to sample 10 domains of study behaviour, and responses to these items were subjected to an oblique multiple group method of factor analysis. The 10 domains of study behaviour were confirmed although there was some significant intercorrelation between domains. In the Q‐analysis, four students were chosen at random from the Ability (high, medium, low) × Faculty (Arts, Science) × Sex matrix, replicated once. These categories were called ‘tests,’ and the items ‘people.’ Using an oblique Procrustes method of factoring, three Q‐scales were obtained. Factor scores on each of the 10 R and three Q scales were obtained for each student, and relationships with sex, faculty and performa