The use of the social model of psychiatry as a form of treatment is increasingly accepted as a legitimate therapy. This is expressed in the continued growth of the 'therapeutic community' regimes. However, investigations into components of the social model, i.e. attitudes of nursing staff in psychiatric hospitals, have not accompanied this growth. A cross-section of research over the past 25 years has shown an omission of behavioural correlations in attempting to measure staff attitudes. Research into attitudes expressed indicate a correlation between age, education, regime in hospital and occupation and the responses to attitude questionnaires. Without behavioral correlations however, the value of this research is very questionable. A few studies have included behavioral correlations such as perception of staff, discharge rates and length of stay in the community. An example of an area in nursing that attitude studies with behavioral correlations could usefully be employed is in examining the effectiveness of using staff from different cultures in therapeutic communities.