The Short NART: Utility in a memory disorders clinic

Beardsall & Brayne (1990) have introduced a method for estimating full‐length National Adult Reading Test (NART) scores using the scores on the first half of the test only. They suggest that this is appropriate for subjects who are of low reading ability and might otherwise find testing distressing. Crawford, Parker, Allan, Jack & Morrison (1991) have subsequently explored the accuracy with which Short NART predicted full NART scores and in addition WAIS IQ scores in a large cross‐validation sample. They concluded that the Short NART could be used with modest confidence when estimating premorbid IQ. However, when applied to 202 consecutive referrals to this Memory Disorders Clinic, the accuracy with which the Short NART predicted full NART error scores was less satisfactory. Results indicated that discrepancies between Short NART and full NART error scores were outside the bounds of both clinical and statistical acceptability. Examination of these results revealed that one possible source of difficulty lay in variation in the accuracy with which words are pronounced. It is concluded that, despite the appeal of a shortened version of the NART to estimate premorbid IQ, without further modification its use in clinical practice cannot be recommended.