Attitudes to and knowledge about elderly people were assessed in 1091 students and lecturers from the London Hospital Medical College (LHMC), London, UK and the English and Computer Science departments of Queen Mary and Westfield College (QMW), London, UK. General knowledge about elderly people was measured by the Palmore 'Facts about Aging Quiz 1'. Attitudes towards elderly people were measured by the Rosencranz and McNevin Semantic Differential scale. A higher level of knowledge about elderly people was found both in medical students and in medical lecturers compared to their counterparts in the English and Computer Science departments (P < 0.001). The cross-sectional data indicated that medical students developed a significantly increasing knowledge about elderly people as they progressed through their training, in contrast to students of English and Computer Science. Scores on the Rosencranz & McNevin scale indicated that attitudes towards elderly people across all three groups of students and lecturers were similar. The scores obtained for all groups indicated that they held approximately neutral attitudes towards elderly people. A significant correlation (P < 0.001) was found between high levels of knowledge about elderly people, and positive attitudes towards them. The necessity of interventions to improve general attitudes towards and knowledge of elderly people among medical students is questioned. It is suggested that future research should look beyond surveys of general attitudes towards elderly people for the causes of the current lack of interest in geriatric medicine.

Objectives. To evaluate the psychometric properties of the 50‐item Communication Style Measure of Norton (1978), to assess the number of its subscales, and to provide normative descriptions of the subscales for providing feedback in teaching sessions. Design. Questionnaire administered to 245 undergraduate medical students and 267 general practitioners. Methods. Confirmatory and exploratory factor analysis of the structure of the 50‐item questionnaire. Regression analysis of scores on subscales in relation to age and sex. Results. Norton's original 10‐factor structure for the questionnaire could not be sustained. Exploratory factor analysis suggested three factors which we labelled effective, dominant and non‐verbal, which were reliable (Cronbach's alpha = .79, .76 and .74), and which each showed different patterns of association with age and sex. On the basis of the factor analysis we also report a brief, easy to score questionnaire with 18 items, and we give normative distributions which might be useful for teaching in small groups. Conclusions. Doctors and medical students found the Communication Style Measure easy to complete and to provide useful normative feedback on their self‐perceptions relative to those of other doctors and students. The 10 statistical factors proposed by Norton are not supported by the data, but three factors are readily identifiable.