Structural efficiency of wais subtests

Abstract

This study investigated the utility of the Cornell Medical Index when used with three types of VA Ss: N P patients, medical and surgical patients (GM&S), and domiciliary members (D). A brief description was given of the domiciliary as an institution and of the characteristics of its membership. Item transparency, response sets, and /or situational factors were seen as possible sources of distortion. On the basis of research findings and clinical experience, certain predictions were made regarding how these three groups should respond to the CMI with a research "set" and in an evaluation situation with implicit or explicit differential action being an outcome.In a research set sample, the relationship generally held that N P > GM&S > D, five out of six comparisons being statistically significant. However, the D Ss failed to score more deviantly than GM&S Ss as was predicted.A file search yielded Ss in each category who had been referred for evaluation for differential action. Each type of S was compared to his counterpart in both the evaluation and research set sample. Evaluation GM&S and N P Ss scored lower than research set GM&S and NP Ss as was predicted. Evaluation D 5's scored higher than research set D Ss; no prediction was made here because of confounded motivations. Both the research set and evaluation samples had frequency distributions considerably displaced toward higher scores than other samples reported in the literature.The results lead us to conclude that the suitability of the CMI as a screening device reasonably free from situational and motivational factors is seriously questioned. However, if this reasoning is not accepted by the reader, the minimal conclusion is that cutting scores reported in the literature should be used with caution and a strong need for local norms is indicated. REFERENCES 1. 2. FOQEL, E. J. Swepsten, E. R., et al. Problems of the aging: Conclusions derived from two years of inhrdisciphary study of domiciliary members in a Veterans Administration Center. Amer. J . Psychiat., 1956, 118 724-730. 3. LAWTON, M. P. ! h e screening value of the Cornell Medical Index. J . consult. Psychol., 1959, g.9, 352-356. 4. MOSLINQ, J. The influence of situational and interpersonal variables in projective testing. P clwl. Bull., 1960, 67, 65-85. 5. ?eterans Adminiitration. Veterans i n V A Domiciliaries. A profile study. PROBLEMThis study evaluates whether Information, Digit Span, Arithmetic, and Picture Completion might represent acceptable samplings of overall efficiency of the WAIS Full scale. "Efficiency" can denote whether a test consists of items required to perform a specific job of measurement.An inefficient test could include items which probe knowledge more representative of some subcultural settings than others, are subject to ambiguous scoring procedures, or are otherwise influenced by extraneous factors, not permitting a stable 'Now at NP, USPHS Outpatient Clinic, Washington, D. C.

Intrasubtest scatter on the WAIS—R as a pathognomonic sign of brain injury.

“…The breadth of relevant literature may be summarized as follows: The relationship between item difficulty and order of item placement in subtests of the 1955 WAIS was such that intratest scatter was relatively uncommon in the protocols of normal subjects (Fink & Schontz, 1958; Payne & Lehmann, 1966). Scattered patterns of success and failure were found in the WAIS protocols of schizophrenics (Nickols, 1963), and similar magnitudes of intrasubtest variability were associated with cerebral lesions (Watson, 1965). Subtests most likely to show scatter were Similarities, Picture Completion, Information, Comprehension, and Vocabulary.…”

“…Scattered patterns of success and failure were found in the WAIS protocols of schizophrenics (Nickols, 1963), and similar magnitudes of intrasubtest variability were associated with cerebral lesions (Watson, 1965).…”

“…Subtests most likely to show scatter were Similarities, Picture Completion, Information, Comprehension, and Vocabulary. Block Design, Digit Span, Picture Arrangement, and Object Assembly were less sensitive (Nickols, 1963). Scatter on the WAIS Vocabulary subtest was found to reliably differentiate brain-injured patients from normal controls, patients with spinal cord injuries, and those with mixed psychiatric diagnoses (Hallenbeck, Fink, & Grossman, 1965).…”

Research with the Wechsler Intelligence Scales for Adults.

“…He found no group differences in intrasubtest scatter on any subtest using two different measures of scatter, but felt that unmatched groups might have differed in this respect. Studying the structural efficiency of WAIS subtests in terms of the percentage of schizophrenics failing a specific item but passing subsequent items in the subtest, Nickols (1963) found the most efficient to be A, BD, D, and OA, that is, they had the least intrasubtest scatter in schizophrenics. Low efficiency or high scatter subtests were S, PC, I, and V. He suggested that the high efficiency subtests were the fairest criteria for matching normals and schizophrenics since irrelevant or fortuitous factors contribute more to the inefficient subtest scores.…”

WAIS-R item difficulties with psychiatric inpatients

“…A relatively uninterrupted series of correct responses is expected in normal populations until the point at which the examinee reaches a ceiling level of ability (Lezak, 1983; Mittenberg, Thompson, Schwartz, Ryan, & Levitt, 1991; Ryan, Paolo, & Brungardt, 1989). Thus, a relatively consistent sequence of incorrect responses is expected, with departures from this expected sequence of successes and failures suggestive of psychopathology (Broder & Oresick, 1987;Feinberg & McIlvried, 1991; Kellerrnan & Burry, 1991; Nikols, 1963) or brain dysfunction (Kaplan, Fein, Morris, & Delis, 1991; Lezak, 1983; Mittenberg, Hammeke, & Rao, 1989; Mittenberg et al, 1991). Previous research has examined the performance of clinical groups on individual items of selected Wechsler subtests, such as Picture Arrangement (Broder & Oresick, 1987) or Information (Norman & Wilensky, 1961) and has found that obtained rank orders of item difficulties do not correspond to those reported for the standardization group.…”