Little evidence for links between memory complaints and memory performance in very old age: longitudinal analyses from the Berlin Aging Study.

Cross-sectional and longitudinal relationships between memory complaint and memory performance were examined in a sample of old-old participants from the Berlin Aging Study (BASE; N = 504, ages 70 to 100, age M = 84.7 at study onset). Participants were measured 4 times over the course of 6 years. Similar to many previous studies, initial cross-sectional memory complaints were predicted by depression and neuroticism, but not memory performance. Subjective age also predicted memory complaint independent of other variables. Latent growth curve models based on age and time in the study revealed that memory complaints did not change in level with age or time, and manifested no reliable random effects (individual differences in change). These models also detected no significant relationship between changes in memory and either initial memory complaint or changes in memory complaint over age or over time. None of the covariates that predicted initial memory complaints were related to changes in memory complaints over time. An autoregressive latent variable model for memory complaints, consistent with a conceptualization of complaints as judgments constructed from beliefs and other influences in the moment, did detect a concurrent effect of memory on memory complaints at the third occasion, controlling on initial complaints. These results suggest that for the oldest-old, changes in memory complaints may not primarily reflect monitoring of actual age-related memory changes, but rather are affected by other variables, including age-based memory stereotypes, neuroticism, depression, and concerns about aging.