THE NEED to develop a systematic way to describe the social field of wards in which experiments in psychiatric treatment are carried out is becoming evident as additional studies are reported. As evidence of this need, two broad reviews of studies of the effectiveness of drug treatments of psychiatric disorder reveal considerable variation in the percentage of improved patients, particularly in the case of antidepressant drugs.2,3 Though a large part of the variation between the findings of one study and another may be attributed to factors such as different experimental designs, different populations, and different techniques for assessing change, there is also the possibility that many studies do not take differences in treatment settings into account. Therefore, it is often difficult to know in what setting good treatment results may be expected. However, to study the impact of different treatment settings on outcome requires codification and measurement of