R.H.B. Cochrane

Construct validity of the National Adult Reading Test: a factor analytic study

Abstract Factor analysis was carried out on the National Adult Reading Test (NART) and Wechsler Adult Intelligence Scale (WAIS) performance of 139 normal subjects. The NART loaded very highly on factor I extracted by principal components analysis (PCA), suggesting that the NART has high construct validity as a measure of general intelligence. Varimax rotation of the three factors extracted by PCA produced a factor structure that was consistent with previous factor analytic studies of the WAIS. The NART loaded highly on Factor I (verbal intelligence) but not on Factor II (non-verbal intelligence) or Factor III (attention/concentration).

Performance on tests of frontal lobe function reflect general intellectual ability

Recent studies have indicated that performance on tests of frontal lobe function are highly associated with general intellectual ability (g). Some authors have even claimed that the available evidence does not support a more specific account of frontal lobe function than to provide a general intellectual function for the performance of goal directed tasks. We examined the relationship between performance on the WAIS-R (as a measure of g) and performance on standard tests of frontal lobe function in 123 healthy individuals. Our results demonstrate that in healthy individuals (i) performance on the most popular tests of frontal lobe function shares significant variance, and (ii) a large proportion of that shared variance is highly associated with performance on the Wechsler Adult Intelligence Scales-Revised (WAIS-R), so that the tests are similar to the extent that they measure g. Performance on the Modified Card Sorting Test (MCST), however, is not related to g. The results support the claim that many tests of frontal lobe function measure primarily a non-specific intellectual function but also indicate that some tests, like the MCST, may be assessing more specific cognitive operations.

Estimation of premorbid intelligence: combining psychometric and demographic approaches improves predictive accuracy

Abstract Regression equations have been developed to estimate premorbid WAIS IQ from (a) psychometric tests which are relatively resistant to the effects of cerebral dysfunction, and (b) demographic variables (e.g. education, occupation). The purpose of the present study was to determine whether regression equations based on the combination of psychometric and demographic variables would account for more WAIS IQ variance than either of the two methods alone. Normal subjects (n = 151) were administered the WAIS and the National Adult Reading Test and had their demographic details recorded (age, sex, education, and occupation). Equations incorporating the NART and demographic variables accounted for 73, 78 and 39% of the variance in Full Scale, Verbal, and Performance IQ. The corresponding figures for equations derived from the NART alone or demographics alone were 66, 72 and 33%, and 50, 50, and 30% respectively.

ESTIMATING PREMORBID INTELLIGENCE BY COMBINING THE NART AND DEMOGRAPHIC VARIABLES: AN EXAMINATION OF THE NART STANDARDISATION SAMPLE AND SUPPLEMENTARY EQUATIONS

Abstract There are two main approaches to the estimation of premorbid intelligence. One of these uses an oral reading test, the National Adult Reading Test (NART). The other estimates premorbid intelligence from demographic variables. It was recently reported that combining the NART and demographic variables in a multiple regression equation provides more accurate estimation of IQ than is afforded by either method alone. The present study was undertaken to determine if this finding would hold in another sample. The sample employed ( n = 120) was that used to originally standardise the NART against the WAIS. Combining demographic variables (age, sex, social class) with the NART significantly increased predicted variance over use of the NART alone. The NART standardisation sample had been administered a short-form of the WAIS. This sample was combined with additional subjects ( n = 151) to produce new equations for the estimation of premorbid IQ. The equations permit clinicians to estimate a clients premorbid intelligence with the NART/demographic method when time pressures have necessitated the use of a short-form WAIS as the current IQ measure.

DEMENTIA IN IDIOPATHIC PARKINSON'S DISEASE : PREVALENCE AND RELATIONSHIP WITH SYMPTOMS AND SIGNS OF PARKINSONISM

A whole population cohort of 157 patients with idiopathic Parkinsonism, most of whom had previously been clinically examined by Mutch (1986a), were assessed to determine prevalence figures for dementia and examine the relationship between dementia, cognitive impairment and Parkinsonian signs. Dementia according to DSM-III-R criteria was diagnosed in 23.3% of all patients (95% confidence interval: 17.1 to 32.4%). Dementia and cognitive impairment were associated with overall measures of Parkinsonian impairment and rigidity, but not tremor, even after controlling for age, sex and education.

Premorbid IQ estimates obtained by combining the NART and demographic variables: Construct validity

Abstract Crowford, Stewart, Parker, Besson & Cochrane ( Personality and Individual Differences , 10 , 793–796, 1989a) have recently built regression equations to estimate premorbid IQ from the National Adult Reading Test (Nelson, 1982) and demographic variables. The construct validity of this IQ estimate was examined in a sample of 151 non-clinical subjects. Factor analysis was performed on WAIS subtests and the NART/demographic estimate (NDE). The loadings of these variables on the first unrotated principal component (which represents g or general intelligence) was examined. The g loading of the NDE was 0.90, which was higher than any of the WAIS subtests. This result suggests that the NDE has very high construct validity and is superior to the NART alone as a measure of premorbid g .

WAIS vs WAIS-R in matched U.K. samples

Abstract A matched samples design was employed to compare the British versions of the WAIS (Wechsler, 1955; Saville, 1971) and WAIS-R (Wechsler, 1981; Lea, 1986) in terms of test difficulty. The WAIS yielded significantly higher IQs than the WAIS-R. The mean difference for Full Scale IQ was 8.2 IQ points, i.e. over half a standard deviation. Encouragingly for U.K. users of the Wechsler Scales, additional analysis suggested that the performance of the WAIS-R sample was consistent with an estimated U.K. population mean ∼ 100.