Date of Award

1994

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

First Advisor

William Drew

Abstract

A combined demographic and present abilities approach to estimate premorbid intellectual functioning was developed and cross validated on unimpaired and closed head injured subjects. The non-clinical sample included 75 non-neurologically impaired individuals divided into two groups. The development sample (n = 50) was used to generate linear regression equations to estimate WAIS-R IQs from the estimated Barona IQ score (Barona, Reynolds, & Chastain, 1984) plus error score on the National Adult Reading Test (NART; Nelson, 1982). The cross-validation sample (n = 25) were individually matched to the clinical subjects on age, education, gender, and race variables. The clinical sample included 25 severe closed head injured (CHI) patients within one year post injury. First, NART performance was shown to be a valid present abilities measure for the estimation of premorbid intelligence. NART performance was shown to be stable in patients with severe head injury. Correlations between obtained WAIS-R IQs and estimated NART IQs (Ryan & Paolo, 1992) were.84,.82, and.75 for FSIQ, VIQ, and PIQ, respectively. Mean NART error scores and estimated NART IQs were the same for CHI patients and matched controls, while mean obtained WAIS-R IQs were significantly lower for the CHI group. The discrepancy between obtained WAIS-R IQs and estimated NART IQs was significantly larger for CHI subjects than for matched normal controls. Next, regression equations to estimate WAIS-R IQs were developed by combining a stable measure of performance (NART error score) with the Barona et al. (1984) demographic estimation of WAIS-R IQs. The variance accounted for by the combined NART-Barona regression equations was 74.39% for FSIQ, 75.90% for VIQ, and 57.19% for PIQ. Standard errors of estimate were 8.56, 8.39, and 10.34 for FSIQ, VIQ, and PIQ, respectively. For the normal cross-validation sample, the correlations between obtained WAIS-R IQs and the NART-Barona estimated IQs ranged from.76 to.87. Estimated NART-Barona IQs were similar for CHI patients and matched controls. The discrepancy between estimated IQs and obtained WAIS-R IQs was significantly less for the combined NART-Barona method than for the Barona et al. (1984) estimation equations.

Pages

94

Share

COinS