Individual-differences assessment of the relationship between change in and initial level of adult cognitive functioning.

Methodological factors may have been partially responsible for the inconsistency in findings from previous investigations into the relationship between change in, and initial level of, adult cognitive functioning. An alternative data-analytic procedure is proposed and applied to data from 277 men (25-76 years of age at initial testing) over 3 measurement occasions (interwave intervals of 6.7 years). Performances on both the Benton Revised Visual Retention Test (BVRT) and the Wechsler Adult Intelligence Scale (WAIS) Vocabulary subtest were analyzed. The findings demonstrate that the significant negative relationships between change and initial status, found on both tests, were partially a function of measurement error. Once adjustments were made for errors of measurement, the previously significant negative relationship for the BVRT data became significantly positive, whereas for the WAIS the relationship remained significantly negative, although to a lesser degree. The importance of accounting for errors of measurement is discussed.