Patient satisfaction can be a useful marker in evaluating the quality of psychiatric care. However, this form of measurement has been hampered by the lack of attention paid to the psychometric properties of instruments devised. The Psychiatric Care Satisfaction Questionnaire (PCSQ) was developed and tests of acceptability, validity and reliability were undertaken using 52 inpatients. Content validity was assessed by surveying psychiatrists, other mental health professionals, MIND workers and patient groups. Thirty-six (67.9%) of all mental health professionals felt that the questionnaire was a useful measure of patients' satisfaction with their care, and many comments also suggested high consensual validity. In terms of content validity, responses from patient groups indicated that the majority (63.6%) felt it to be a useful measure of patient satisfaction and 70.5% felt all the areas covered were important. Concurrent and criterion validity were also good, with high correlations with existing scales and with those on a section being significantly less satisfied on the PCSQ (F = 13.3, P = 0.0004). Test-retest reliability was adequate (Cohen's kappa 0.48-0.80) and the internal consistency of the PCSQ was good, at 0.82 (Cronbach's alpha). The PCSQ performed well in all aspects of validity and reliability, indicating that it has good psychometric properties and can be a useful tool for measuring patient satisfaction. However, there remains the need to analyse more closely the concept of satisfaction and its determinants.