Twenty-six per cent of patients in two Inner London districts were admitted to acute wards under the provisions of the Mental Health Act. Compared with those not under compulsion, they were young, male, more likely to be of black Caribbean origin, and to have a diagnosis of schizophrenia of short duration. The hypothesis is tested that ethnicity determines rates of compulsory admission independently of the other factors.

Sampling and data collection methods were described in the first paper. Statistical analyses included a log-linear analysis of six key variables: compulsory admission, challenging behaviour, diagnosis, ethnicity, age, and sex.

There were no substantial differences between districts. Analysis provided two similar statistical models. In both, admission under the Act was strongly associated with challenging behaviour and diagnosis of schizophrenia. In the model of best fit there was no significant interaction term for ethnicity and compulsion. In the second model there was a weak association.

Ethnicity did not appear to be of outstanding importance in decisions to use the Mental Health Act. There was a strong link between ethnicity and diagnosis, independent of compulsion. Differences between the districts made no major contribution to the rates of compulsory admission.