Relationship Between Actual Intelligence & the CAMCOG-R in Cognitively Normal Elders

We examined (1) the relationship between actual intelligence and the CAMCOG-R performance in cognitively normal elders and (2) the presumed impact of age and education on the relationship. Our retrospective study of 94 participants assessed in a specialized setting, measuring actual intelligence (general and specific) with the Wechsler Adult Intelligence Scale-III, the Full-Scale Intelligence Quotient (WAIS-III FSIQ), and the specific subtests thereof. The CAMCOG-R correlated largely with FSIQ (.629, p < .001; .708 when controlled for age). The CAMCOG-R's scores increased across three different IQ levels as IQ increased. FSIQ accounted for 40% of the CAMCOG-R's variability. Combined, FSIQ and age accounted for 59% of the CAMCOG-R's variability. FSIQ had a stronger weight than age (beta: .644 and -.439, respectively). Regression analyses with FSIQ's subtests, age and education, found the best-equilibrated model to include Information, age, and Picture Completion. Combined, they accounted for 55% of the CAMCOG-R's variability. Information received the strongest weight (beta: .411). Overall, education was nonsignificant. Thus, a significant relationship existed primarily between general intelligence and the CAMCOG-R and, to a lesser extent, between the above-mentioned subtests of intelligence and the CAMCOG-R. Adding age increased the prediction rate. Actual intelligence and age should be considered when interpreting the CAMCOG-R's scores. The practical implementation of the intelligence factor is complex.