The training and generalization of conversation behaviours in psychiatric in-patients: a controlled study employing multiple measures across settings

Abstract

This report provides an account of a conversation training programme for psychiatric in-patients, developed on the basis of previous work in the area and on analyses of normal conversations carried out by the author. A further aim was to generalize these increases in social behaviour to patients' general ward settings. Generalization procedures include tasks for patients to do between training sessions and a reduction in the number of stimulus differences between treatment and generalization settings. Two control groups were employed to assess the effectiveness of this treatment as compared to (1) an existing behavioural approach to the problem of withdrawn patients and (2) a group who received similar increases in staff attention, tangible reinforcement, and who also had their own attention focused on social behaviour. This latter control group then went on to receive the training given to the experimental group. Patients were assessed before treatment, after treatment and at 8-week follow-up. Measures were taken on frequency of conversation during treatment, frequency of conversation on the ward, quality of various social behaviours during videotaped interviews and quality of social behaviours during audiotaped conversations. The results indicate the following: (1) both behavioural treatments increased patients' utterances during treatment sessions; (2) the experimental treatment was superior in generalizing this increased frequency, although these increases reduced at follow-up; (3) the experimental group show some superiority when assessed on videotape, and gains made in the quality of interaction do not reverse as markedly as in frequency; (4) on the audiotaped conversations, the experimental group improved their social ability considerably and these improvements were still evident at follow-up. The behavioural treatment control showed some improvements on generalized measures but not as much or as consistently as the experimental group. The attention control group went on to replicate the experimental group treatments and results.