9 - The Accuracy of Formal Neuropsychological Methods

Abstract

Accuracy is equivalent to the rate of error and is a primary component of predictive validation. The criteria for validity and accuracy have not been adequately examined in neuropsychology. The accuracy of a test procedure cannot be better than its criterion, which may not be above 80% in neurology. This chapter presents a a crucial discussion of the accuracy of various formal methods that have been used in neuropsychology. It attempts to be as thorough as possible in its selection of procedures to examine. Nevertheless, this examination is restricted to procedures that have developed from the Halstead–Reitan Battery (HRB). Various types of psychometric procedures that are used to determine the accuracy of brain-damage studies include the Halstead Index (HI), the average impairment rating (AIR), the Neuropsychological Deficit Scale (NDS), and the average impairment scale (AIS). Accuracy is another way of stating the rate of error for a procedure. The major examined procedures are indexes, decision-tree algorithms, discriminate analysis, and clinical judgment. The chapter also deals with the criticism that neuropsychologists have applied to these formal methods. Finally, it briefly discusses some of the newer methods of creating formal procedures