Abstract As a self-report questionnaire, the Cognitive Failures Questionnaire (CFQ) was originally devised to measure perception, memory, and motor lapses in daily life. CFO scores have been found to correlate with some psychiatric symptoms associated with stress; hence, high scores on the CFQ are considered by some as an indicator of increased vulnerability to stress. Attempts to identify a stable factor structure for the CFQ. have produced disparate results. However, there is a measure of agreement with regard to the presence of a “general cognitive” factor that includes loadings from most items and accounts for the lion's share of the variance. Not enough is known about the performance of the CFQ in clinical populations to use it as a measure of change. The current study sought to explore the performance of the CFQ in three groups of patients, organic (n = 209), mixed (n = 115), and functional (n = 322), and to identify correlations with measures of psychiatric morbidity (General Health Questionnaire [GHQ]), depression (Beck Depression Inventory [BDI]), and recognition memory (Signal Detection Memory Test). In the organic and functional samples, the CFQ score significantly correlated with the BDI and GHQ but not with the recognition memory measure. Three factors were found to be common to the organic and functional samples: cognitive, dissociation, and clumsiness. No characteristic pattern of CFQ item endorsement to differentiate between the organic and functional samples was found. Seven items of the CFQ performed badly because of ceiling or floor effects. The “negative” results reported herein are of relevance to researchers who may be planning to use the CFQ in clinical research. The CFQ remains a promising instrument, particularly on account of its “ecological” features, but far more investigation is needed before it is used as a standard measure in clinical practice.