Administered the Halstead‐Reitan neuropsychological test battery to a heterogeneous sample of 52 brain‐impaired patients and 202 non‐impaired college and community volunteers. The volunteers were assigned randomly to either the control group or to one of four faking groups, which differed only in terms of type of brain damage Ss were to fake. The Right and Left groups were told to fake unilateral damage to only one hemisphere, the Diffuse group was told to fake damage to both hemispheres, and the Nonspecific group simply was told to fake brain damage. The author achieved a hit rate of 94.4% on subjective classification of a subsample of 195 Ss into brain‐impaired vs. non‐impaired categories. Stepwise discriminant analysis of the entire sample yielded two functions that achieved hit rates of from 94.4% to 97.2% for various base rates of malingering. Discrimination between control and faking Ss was much less accurate, and the latter were highly unsuccessful at generating believable patterns of lateralized cortical impairment. Posttest interviews were conducted to obtain information concerning faking strategy as well as factors that inhibited or facilitated the efforts to fake.

The Satz-Mogel abbreviated Wechsler Adult Intelligence Scale (WAIS), while yielding high correlations with the standard WAIS scales, has been criticized for introducing sufficient subtest unreliability to prohibit profile interpretations. Using multivariate profile analytic techniques (Cattell's r p and hierarchical grouping analysis) and sampling from both brain-injured and psychiatric populations, these forms were compared in terms of overall pattern similarity and actuarial classification agreement. For both populations, the results provide strong evidence that the Satz-Mogel abbreviated WAIS may be substituted for the standard WAIS for both general IQ assessment (in agreement with previous research) and global profile analysis (contrary to previous research). Further, the argument was advanced that the multivariate techniques of the present study provide a statistically more reliable inferential base for WAIS profile analysis than does the comparison of individual subtests.

Hypnotic screening of 108 subjects was carried out in groups of 4 or fewer using the Harvard Group Scale. Attempts to influence susceptibility scores (two objective observer scorings and two self-scores) via experimenter bias (role playing combined with experimenter expectancy), experimenter status, and subject belief (information to subjects about their likelihood of being hypnotized) failed to yield statistical significance, except in one case. Experimenter bias was significant at the .05 level in only one of the four three-way AJTOVAs needed to analyze the four duplicate susceptibility scorings. This result accords with an insufficiency of evidence for important experimenter bias effects in hypnotic research. As such, it agrees with a recent critical review of Roscnthal's work as it applies to any well-structured, professionally conducted psychological experiments. The research significance of any very small experimenter bias effect that may have occurred in this study is discussed in light of the minute strength of relationship, or percentage of variance accounted for by each source of variation, in each of the four ANOVAs.

The Mooney Closure Faces Test has been employed in several studies of right temporal-lobe function. However, the information provided by these studies is somewhat restricted because there has been no systematic attempt to determine what constitutes “normal” performance on this instrument. The present study attempts to rectify partially this situation by testing samples of three distinct populations (college students, vocational-technical students, and relatives of indigent patients) of neurologically intact subjects who differed in age, IQ, educational level, sex, and handedness. Regression showed all of these variables to be significantly correlated with performance on the Mooney Closure Faces Test. The final linear regression model was correlated 0.718 with the Mooney scores and resulted in predictions which were approximately 30% more accurate than those made using the mean of the entire sample alone.

No abstract