No abstract

Ten years of validity research concerning the Slosson Intelligence Test (SIT) is reviewed. The equivalence of the SIT to the Stanford-Binet and the Wechsler Scales is examined and the relationship of the SIT to achievement measures is described. The importance of supplementing correlational findings with regression analysis, mean difference data, and information on misclassification is emphasized by the reviewers. Declsions concerning individuals based on SIT I& scores are viewed as particularly risky in situations involving groups of subjects characterized by a relatively narrow range of I& scores.The last decade has seen an increasing use of the Slosson Intelligence Test (SIT) in a variety of professional and educational settings. In some situations, the SIT is accepted as a valid individual intelligence test and is used for such purposes as special class placement. More and more in the last few years the SIT is being used for retesting purposes where individual intelligence test results are needed or required. In view of these developments, a review and evaluation of validity research concerning the SIT should prove useful t o psychologists, counselors, and educators.The validity of the SIT will be examined within three contexts: (a) comparability or equivalence of the SIT to the Stanford-Binet and Wechsler Scales as an index of general intelligence; (b) usefulness of the SIT in predicting achievement; and (c) comparative usefulness of the SIT for assessment or clinical interpretation purposes. Table 1 presents a summary of SIT-S-B validity results. Of 12 reported correlations, 8 are in the .9Os range. The correlations in the .60 to upper .?Os range all are based on samples of subjects characterized by a restricted range of I& scores (note the standard deviations). The median correlation between the SIT and the S-B is .91 for this group of studies. Thus, the overall correlational picture suggests that with groups of subjects in which the range of scores is not restricted, the SIT relates t o the S-B to such a high degree that one can be confident that subjects are being ranked very nearly in the same order by the two tests. Further, only two of the 14 groups compared in these studies showed more than a two I& point difference between SIT mean IQ and Stanford-Binet mean I&. I n spite of these impressive findings, caution should be used in individual diagnosis and in decision-making in situations where the decisions are in a restricted I& range context, such as identifying an individual as a "slow learner" vs. "mildly retarded," or distinguishing between mild vs. moderate levels of retardation. Even in the above studies where the correlation between the SIT and the StanfordBinet was very high, substantial differences in I& could be observed for a significant percentage of the subjects. For example, in the Stewart, et al., (1971) study, 16% of their subjects had SIT IQs which were 10 or more points different from their Requests for re rints should be sent to Kenneth D. COMPARABILITY OF THE SIT TO THE STANFORD-BINET

AND CONCLUSIONS The Wide-Range Vocabulary Test was selected as an instrument t o provide verbal I& equivalent measures a t several Wisconsin Correctional Institutions.Regression equations were constructed for various age groups between Wide-Range raw scores and Otis IQs. The resultant IQ equivalents were then correlated with Otis IQs as well as with WAIS verbal, performance, and fullscale scores. Moderately high, positive correlations were consistently obtained between these I& measures and the Wide-Range I& equivalents. I n general, the Ioregoing results suggested that the Wide-Range Vocabulary Test represented a valuable addition to our screening battery, providing reliable and valid information about level of verbal intelligence at substantially reduced cost and a t approximately one-third the administration and scoring time of the Otis test. I n a subsequent study, t,he relationship of Wide-Range scores to school achievement as well as the relationship between discrepancies in verbal and performance IQs and deficits in general adjustment and school achievement will be assessed.

No abstract

Stein (1961) recently reported highly successful discrimination of brain damaged subjects from non-brain-damaged subjects by use of a rapid, easily administered, age-corrected, perceptual-motor task, the "Symbol-Gestalt Test." While the test has been used by Stein on a number of samples in prepublication studies, questions may be raised concerning the generalizability of Stein's findings to different populations, examiners, and locales.In the course of a series of studies of brain damage, it was possible to test the applicability of the Symbol-Gestalt Test to two populations drawn from widely separated Veterans Administration hospitals,