Published online by Cambridge University Press: 18 September 2015
A consistent feature in dementia is an overall intellectual deterioration. The present study investigated the intellectual performance of patients with Alzeheimer's disease (AD) and multi-infarct dementia (MID) using the Wechsler Adult Intelligence Scale (WAIS). For reliability and generality purposes two independent samples of patients were collected. Significant differences in Education Level (EDU) and Performance IQ (PIQ) were obtained for the first sample, with the AD group having a significantly higher EDU level. The MID group obtained a higher PIQ. No significant differences were found in the second sample, but EDU level approached significance with the AD group again having a higher EDU level. A discriminant function analysis classified 81% of the patients in the first sample and 100% diagnostic accuracy was obtained for the second sample using the II predictor variables. A maximum R stepwise regression was performed in order to detect the “best” model of variables discriminating between the diagnostic groups. For the first sample the “best” model was the two variable model, including EDU and Full Scale IQ, accounting for 40% of the variance. The simplest model for the second sample was the one variable model including EDU, accounting for 20% of the variance. Quantitative, differences were found between the AD samples. Time since onset of the disease was offered as a possible influence in the quantitative differences in the AD samples. Sampling biases in the behavioral study of dementia are discussed. Successful replication was obtained.