A WAIS-S short form for the assessment of intelligence as a confounding variable is proposed. A number of 161 complete WAIS-S protocols, relating to women of lower-middle and low socioeconomic status were analysed. Six subtests (Social Comprehension, Similarities, Vocabulary, Block Design, Picture Completion and Object Assembly) were chosen using two criteria: specific functions evaluated by each subtest and correlation coefficients relating IQ to the subtests. Five WAIS-S short forms were evaluated: one consisting of six subtests and the others with different combinations of four of them. Data showed that they were all appropriate for the measurement of intelligence as a confounding variable; the correlation coefficients between short forms and total IQ ranged from 0.90 to 0.94. The standard errors of the short forms for the estimation of total IQ were of 3 or 4 points, and 3 to 7% of the subjects were wrongly classified. When maternal intelligence had to be measured in relation to infant development the short form that included Social Comprehension, Similarities, Block Design and Picture Completion seemed the most appropriate. Nevertheless when choosing a WAIS-S short form not only economy of time but also the specific relevant functions assessed in each subtest must be considered.
Wechsler scales; Intelligence; Confusion; Socioeconomic factors