Abstract
The effects of a five-day in-service training in behaviour therapy for psychiatric nurses were assessed by naturalistic observations in three NHS long-stay wards. The observations focused on nurses'' and patients'' independent and interactive behaviors as a function of the course of training, during an unstructured part of the ward programme. A comparison of the behaviourally trained nurses and their colleagues indicated that the training had produced no systematic effect on independent activities and was associated with significantly less frequent or appropriate interactions with patients. These findings were contrasted with the nurses'' proficiency in their structured use of behavioural techniques and led to an interpretation of patient control over their use of these new skills in default of staff management procedures. More broadly, the findings were regarded as an elaboration of the respective roles of staff training and management in institutional change. They were also viewed as evidence that circumscribed and short-term evaluations may be contradicted by more ''ecological'' evaluations, which unobtrusively observe the broader consequences of our interventions into systems. In the long term these consequences may ultimately determine the effectiveness of our training endeavours and hence there is a pressing need for more process evaluation of the present kind.