AbstractThe reliability and predictive validity of a short form of Honey and Mumford's Learning Styles Questionnaire were examined. Students' learning preferences were measured by self‐report ratings on their degree of likings for 19 teaching and learning activities commonly used in higher and professional education. Subjects were 381 second‐year undergraduates enrolled in accountancy, engineering and communication programmes at the Hong Kong Polytechnic. The alpha coefficients for the four learning styles scales were quite low, ranging from 0.311 for the Pragmatist scale to 0.421 for the Reflector scale. Factor analysis of the items did not reveal any coherent factor structure congruent with the underlying constructs. However, significant though weak correlations were found between the learning styles scores and the learning preferences of the students as hypothesi