Development, Psychometric Properties, and Validity of the Hopkins Adult Reading Test (HART)

Abstract

We describe the development of a 35-item, oral word-reading test with two equivalent forms (HART-A and HART-B) designed to estimate premorbid abilities. Both forms show excellent internal consistency (coefficients alpha>.91) and test-retest reliability (Pearson rs >.90). HART performance was combined with demographic variables to generate regression equations that predict IQ scores obtained concurrently and 4-8 years earlier. The resulting models explained 61% of full scale IQ (FSIQ) variability in 327 healthy adults. The FSIQs that can be estimated range from below 73 to above 131. Combined with demographic variables, these two brief word reading tests accurately predict a broader range of IQs than Blair and Spreen's (1989) longer version. Equivalent forms make it especially useful for longitudinal studies.