Two studies are described here. In the first, an attempt is made to examine the relationship between the complaints of anxiety and depression using the results of a survey of a dermatology clinic. The Hospital Anxiety and Depression scale (HAD) and Clinical Interview Schedule (CIS) were both used in 117 unselected clinic attenders. The results showed that the self-report scales (HAD and first section of CIS) showed a closer relationship between anxiety and depression than the second section of the CIS that requires clinical judgement. In the second study, an attempt was made to distinguish between two possible explanations for this phenomenon; that psychiatrists use the constructs of anxiety and depression more carefully than patients, or alternatively that psychiatrists show an observer bias. A sample of psychiatrists were asked to complete symptom profiles for anxiety and depression; both for an imaginary 'patient' and for their own emotions, allowing psychiatrists' beliefs concerning the relationship between anxiety and depression to be studied. The results indicated that the psychiatrists thought there was a negative correlation between anxiety and depression in patients in contrast to the consistent empirical finding that anxiety and depression are positively correlated. Psychiatrists also thought anxiety and depression were less closely related in patients than in themselves. It is likely that psychiatric assessments of anxiety and depression requiring clinical judgement on the part of the interviewer are subject to observer bias. It is argued that self-report measures of psychiatric symptoms will provide more convincing evidence to resolve diagnostic disputes.