Abstract
The composition of abbreviated forms of the Stanford-Binet, WISC, WPPSI (and briefly the WAIS) are described. Methodological problems in constructing, validating, and using short forms are presented and evaluated. Short forms are seen as practical and some judged to be valid for certain populations in predicting full scale intelligence. However, traditional methods of devising short forms are generally judged to be inadequate. Multiple correlation techniques which retain only certain subtests drastically reduce the flavor of the full scale, and arbitrary or random deletion of items from whole scales or subtests does not guarantee retention of the most predictive tasks. Statistically grounded item analysis is offered as a meaningful model to future researchers who seek to construct shortened individual intelligence tests. However, the question of whether the expense in research time is worthwhile in this regard is posed and answered in the negative. Alternate suggestions for evaluation of intelligence that may increase both validity and information per unit of assessment time are given.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have