| Abstract: | Changes in the full scale reliability of the WISC-R were computed at three age levels when each subtest was omitted by itself. Reliability was then determined when combinations of the two to five subtests which independently lowered the full scale reliability the most were omitted. The same procedure was followed with those subtests which independently had the smallest effect in lowering full scale reliability. The deletion of any one subtest had a negligible effect on reliability. Only when the combination of the five subtests having the greatest independent effect on full scale reliability was omitted did the reliability drop below.90. Cautions were rioted concerning the exclusion of sub-tests even when reliability remains acceptably high. |
