Studies of aggressive behaviour in schizophrenia: is there a response bias?

Of 73 patients who met selection criteria to enter into a study on aggressive behaviour in schizophrenia, 11 patients (15.1%) did not participate. The participants and non-participants were similar in age, gender ratio and proportion who had aggressive behaviour. The participants, however, had a longer duration of illness, a longer duration of current admission, were more likely to suffer from residual schizophrenia, but less likely to suffer from disorganized schizophrenia and were less severely ill than the non-participants. These results indicate the need, in studies of aggressive behaviour in schizophrenia, to consider non-response bias as a confounding variable.