Background: Psychology and educational theory has a long tradition of research into learning styles. However, the current educational policy and practice interest in learning styles in the UK has resulted in concepts and practices being adopted with little rigorous empirical evaluation. Purpose: This small-scale, experimental study aimed to test the reliability and validity of an available inventory designed to identify learning styles (visual, auditory and kinaesthetic). Sample, design and methods: Nineteen children, aged 7-10 years, with different styles as identified by the inventory, participated in a class experimental evaluation of their differential response to the teaching of word spelling using different teaching methods. Results: The study found that the visual and auditory scales, but not the kinaesthetic scale, were reliable (internally and re-test). The three groups of pupils with different learning styles--visual only, auditory only and mixed visual and auditory--showed different gains to teaching that matched these styles (visual and auditory teaching approaches). Retention of word spelling was higher one week after the teaching when the teaching matched the learning style. Conclusions: The findings in this exploratory study suggest the significance of learning style for classroom-based teaching. Learning style inventory is appended. (Contains 2 figures, 6 tables and 1 note.)