"The unreliability of IQ tests has been proved by numerous researchers. The scores may vary by as much as 15 points from one test to another,29 while emotional tension, anxiety, and unfamiliarity with the testing process can greatly affect test performance.30 In addition, Gould described the biasing effect that tester attitudes, qualifications, and instructions can have on testing.31 In one study, for example, ninety-nine school psychologists independently scored an IQ test from identical records, and came up with IQs ranging from 63 to 117 for the same person.32

In another study, Ysseldyke et al. examined the extent to which professionals were able to differentiate learning-disabled students from ordinary low achievers by examining patterns of scores on psychometric measures. Subjects were 65 school psychologists, 38 special-education teachers, and a "naive" group of 21 university students enrolled in programs unrelated to education or psychology. Provided with forms containing information on 41 test or subtest scores (including the WISC-R IQ test) of nine school-identified LD students and nine non-LD students, judges were instructed to indicate which students they believed were learning disabled and which were non-learning disabled.33

The school psychologists and special-education teachers were able to differentiate between LD students and low achievers with only 50 percent accuracy. The naive judges, who had never had more than an introductory course in education or psychology, evidenced a 75 percent hit rate.34 When Ysseldyke and Algozzine cite Scriven, they clearly show their belief that the current system is in trouble:

The pessimist says that a 12 ounce glass containing 6 ounces of drink is half empty — the optimist calls it half full. I can't say what I think the pessimist could say about research and practice in special education at this point, but I think the optimist could say that we have a wonderful opportunity to start all over!35 "

from

www.audiblox2000.com/dyslexia_dyslexic/dyslexia014.htm