'Lost and found': some logical and methodological limitations of self-report questionnaires as tools to study cognitive ageing.

Laboratory tests show a decline in memory efficiency with age, but may not reflect highly practiced everyday memory skills. Self-rating questionnaires probe memory efficiency in everyday life, but seven logical and methodological problems ambiguate interpretation of the data they yield. To explore these problems, 442 volunteers aged from 50 to 85 years were given the AH 4 IQ tests, Mill Hill Vocabulary Tests, Beck Depression Inventory (BDI), Broadbent Cognitive Failures Questionnaire (CFQ), Sunderland & Harris Memory Failure Questionnaire (MFQ), the 'Lost and Found' Questionnaire (LF, a new domain-specific probe of the variety, frequency and aetiologies of losses of possessions) and two laboratory memory tests. All self-rating questionnaire scores correlated modestly with each other. Counter-intuitively, individuals over 60 years old reported fewer lapses on the LF and on the CFQ. No SRQ scores correlated with performance on laboratory tasks. CFQ scores did not correlate with IQ test scores but high IQ test scorers reported greater recent decline in memory efficiency (MFQ) and greater current frequency of losses of objects (LF). CFQ and LF scores correlated positively with Beck Depression Inventory scores, suggesting some common loading for poor self-regard. The results suggest that subjective self-ratings cannot reflect absolute levels of everyday competence, but only the relative success of individuals' adaptations to idiosyncratic environments which, as they age, may change faster than their abilities. However, besides exposing the limitations of self-report techniques, these analyses also suggest practical ways of adapting and using SRQs which may allow us to gain better insights into changes and individual differences in functional efficiency in everyday life.