The author proposes a way of maximizing the clinical usefulness of limited normative test data through the use of overlapping cell norm tables. Adjacent cells differ in means, but overlap in range, and any one norm subject may appear in one or more adjacent cells. The neuropsychological test data of 363 nonclinical subjects are used to illustrate the approach.

The Missouri Children's Behavior Checklist is a set of descriptions of children's behavior that may be rated by a child's parent. This checklist is based on the literature dealing with observers' reports of children's behavior in a number of different settings and consists of items that have been shown to cluster together into six relatively independent dimensions or scales (Aggression, Inhibition, Activity Level, Sleep Disturbance, Somatization, and Sociability) . The scales can be rated with reasonable reliability, and the scale scores, based on ratings by mothers, discriminate at least two groups of clinically different children. The dimensions of the Missouri Children's Behavior Checklist appear to be sufficiently independent, reliable, and discriminating to warrant further use as one approach to the quantitative evaluation of the behavior of children.

The aim of this study was to examine whether at least a subgroup of patients with toxic multinodular goiter may have autoimmune thyroid disease. Thyroid-stimulating immunoglobulin (TSI) activity, measured by a sensitive bioassay employing cultured human thyroid cells, was determined in patients with toxic multinodular goiter and other thyroid disorders. All patients with active Graves' disease (n = 47) had detectable serum TSI activity, whereas TSI was undetectable in patients with thyroid disease not believed to be of autoimmune origin: toxic adenoma (n = 13), cold nodule (n = 5), and nontoxic goiter (n = 19), with a single exception in the latter group. Toxic multinodular goiter (n = 26) was diagnosed based on clinical and laboratory evidence of hyperthyroidism associated with a multinodular goiter on palpation and scintiscan. The toxic multinodular goiter group was then subclassified according to scintiscan pattern (type A, diffuse but uneven distribution of technetium uptake; type B, multiple discrete nodules of varying size and function). All but 1 of the 11 TSI-positive toxic multinodular goiter patients had a type A scintiscan pattern. The patients with the type A scintiscan pattern were younger and more often had elevated antithyroid antibody titers, ophthalmopathy, and concurrent development of goiter and hyperthyroidism (rather than long-standing goiter preceding hyperthyroidism) compared to the type B patients. Thus, a subgroup of patients with clinically defined toxic multinodular goiter (type A) probably have autoimmune hyperthyroidism (a variant of Graves' disease), while in another subgroup (type B) hyperthyroidism is not related to an autoimmune etiology (a variant of toxic adenoma).

P R o B L E MSatz and Mogel@) presented an abbreviated form of the WAIS made up of items drawn from each subtest. This resulted in an instrument whose subtests showed high correlations (range .77 to .97) with the original subtests from which they were drawn, and whose Verbal, Performance, and Full Scale IQ's correlated .99, .97, and .99, respectively, with the original IQ's. The abbreviated scale was devised "to represent the complete variety of functions tapped by the original WAIS while still effecting a considerable saving on This author collected siniilar data at about the same time that Satz and Mogel did. This paper presents these results, compares them with those of Satz and Mogel, and gives a further data analysis which leads to some statements of caution. P. 78) Inf.97 .89 Comp .92 .85 k i t h .97 .90 Sim .93 .91 Dig Span 1.00 1.00 VOC .96 .97 Verbal I& .99 . 99 METHOD Subjects.The WAIS records of 35 in-patients were taken from the test files of a municipal psychiatric installation. The sample consisted of 18 women and 17 men, ranging from 17 to 67 years, median age 35, and who represented a variety of psychiatric diagnoses.Procedure. A split-half WAIS was derived by taking successive pairs of items from each subtest and choosing one of them a t random for inclusion.2 The Digit Span and Digit Symbol subtests were left in their original form. This split-half WAIS differs from the form devised by Satz and Mogel in that they used a inore set system for choosing their items (odd items, or every third item) and three of their subtests (Inf., Voc., Pic. Compl.) consisted of a third of the original items.For the split-half WAIS, the raw score for each subtest was multiplied by two, the results converted to scaled scores according to We~hsler's(~) procedure, and IQs obtained in the usual way. RESULTSThe Full Scale IQ's of the original form range from 52-132, median 94. The split-half Full Scale IQls range from 52-132, median 93. Table 1 presents the correlations between the original and abbreviated subtests and scales of the WAIS, TABLE 1. CORRELATIONS BETWEEN ORIGINAL AND ABBREVIATED SUBTESTS AND SCALES Pauker Satz and Mogel Pauker Satz and Mogel Verbal Subtests 7 7 Perf. Subtests t t Dig Sym 1 .oo 1.00 Pic Comp .89 .92 Block Des .94 .84 Pic Arr .91 . 91 Obj Ass .83 .77 Perf. I& .98 .97 Full Scale I& .99 .99 'The data were collected while the writer was on the staff of the Malcolm Bliis Mental Health Center. St. Louis. Missouri., --~---~ ~ -~, -. . *Following are the items making up the split-half WAIS.

No abstract