The presence and nature of intellectual impairment in the affective disorders has always been a matter of controversy. The NIMH Collaborative Depression Study (Clinical Branch) employed the Shipley Institute of Living Scale to assess the IQ and abstraction ability of 954 routine psychiatric inpatient and outpatient admissions for affective disorder and another group of 98 sociodemo-graphically matched controls. Patients were interviewed by trained clinical raters using the SADSRDC, then completed the Shipley during their hospital stay or initial assessment. Patients with affective disorder exhibited significantly lower IQ and abstraction ability scores than controls. Within the patient sample, those with primary recurrent unipolar depression and greater depression severity exhibited the least deficient IQ and abstraction ability scores. Endogenous, psychotic, or retarded subtype depression; mania; schizoaffective disorders; and concomitant alcohol or drug abuse were not correlated with more or less severe intellectual deficits. The finding that clinician ratings of "cognitive" symptoms did not identify patients with greater intellectual or abstraction ability impairment points to an area where syndrome identification is more useful than symptom identification.