Although research has shown that higher levels of within-person variability across cognitive tasks (dispersion) are associated with cognitive decline in clinical samples, little is known about dispersion in comparatively younger, non-clinical, and national samples. A better understanding of dispersion is needed to elucidate for whom and under what circumstances dispersion can be used as a reliable indicator of cognitive health.
We used data from the Midlife in the United States Study (MIDUS; n = 2,229; Mage = 56 years, range = 33-83; 56% female) to: (a) characterize dispersion and its cross-sectional correlates in a non-clinical, adult lifespan sample and (b) examine changes in dispersion over time to determine for whom changes in dispersion may reflect better or worse cognitive aging.
Correlations showed higher levels of dispersion were associated with higher levels of mean performance at both waves (rs = .28-.29). Autoregressive main effect models showed that increases in dispersion were associated with less decline in mean performance over the two-wave, 9-year follow-up period (β = .17, p < .001). Moderation models showed that the link between change in dispersion and change in mean performance was pronounced in comparatively older adults (β = .28), women (β = .27), individuals with less education (β = .23), and those with lower income (β = .23) (all ps < .001).
Findings suggest that increases in dispersion may not always be maladaptive in normative, adult lifespan samples and may reflect healthier cognitive profiles in individuals who are at greater risk for cognitive impairment.