ERIC Number: ED179295
Record Type: Non-Journal
Publication Date: 1979-Sep
Reference Count: N/A
Statistical Weaknesses of a Unitary Construct of Intelligence.
Horn, John L.
A brief discussion of theories of general intelligence precedes a report of relevant empirical data. Results from the factor analysis of more than 20 sets of Wechsler Adult Intelligence Scale (WAIS) data indicate that the WAIS clearly is not a one-factor scale. It does not measure a single, general intelligence. Roughly 17 percent of the reliable variance of the WAIS remains unaccounted for by a three-factor model. There must be at least three, and probably at least four, major components responsible for the individual differences that are represented by the total IQ obtained with the WAIS. Additional analyses on other data suggest that as many as eight major factors are needed to account for the individual differences measured by intelligence tests. General intelligence and the IQ cannot suffice as a basis for a science of human abilities because they have not been defined in terms that are objective enough to ensure that the same phenomena can be observed under different conditions of observation. There is, however, a wide range of phenomena that to some extent can be described by use of a concept of general intelligence and a measure of IQ; IQ measures of general intelligence do provide a crude basis for making diagnoses and decisions. Future studies of intelligence, though, should sample a wide domain of behaviors across the life span and across cultures. Exclusive reliance on a single measure should be avoided. (Author/RH)
Publication Type: Reports - Research; Speeches/Meeting Papers
Education Level: N/A
Authoring Institution: N/A
Identifiers - Assessments and Surveys: Wechsler Adult Intelligence Scale