Describing clinical teachers' characteristics and behaviours using critical incidents and repertory grids.

Context: Completion of a rating questionnaire is the method used most frequently to evaluate a teacher's performance. Questionnaires that largely assess 'high-inference' teaching characteristics, such as 'enthusiasm' and 'friendliness', require the observer to make a judgement about the teacher but do not describe what the teacher actually did and so have limited use in providing feedback. Measures of 'low-inference' teaching behaviours (i.e. those that are concrete and observable), such as frequency, amount or types of verbal interaction, do not demonstrate how these are linked to good teaching.

Objectives: To describe high-inference teacher characteristics and define the associated low-inference behaviours.

Methods: A purposive sample of consultants, postgraduate and undergraduate students, nurse lecture practitioners and patients were selected for semistructured interviews using repertory grids and critical incidents to elicit preferred characteristics and behaviours of clinical teachers. Interviews were audiotaped, transcribed and then content-analysed using a framework to pair teachers' characteristics and their behaviours.

Results: We identified a variety of preferred high-inference characteristics and their associated observable and recordable low-inference behaviours.

Discussion: We carried out a study that included all participants in clinical teaching and found that participants differed in their preferred characteristics and behaviours. It is important for future research to look at behaviours interdependently, rather than alone, and to take into account the evidence that participants tend to infer characteristics rather than think in terms of behaviours. This information will be used to inform the development of a formative tool for evaluating clinical teaching.