The two studies with the lowest levels of attrition
(Milton et al., 1978; Drury et al., 1996) used inpatients.
Participation refusal rates were not uniformly
assessed across studies, even though these statistics
may be more meaningful than attrition rates in studies
of schizophrenia.