Psy 421: Systems and Theories of Psychology Intelligence testing and psychometric approaches to Psychology.

1. Cronbach identified two scientific psychologies: Causal Effects and Individual Differences. Two ways to study psychological phenomena. Classical psychology comes from the first: What is the effect of x (some experiential or historical variable) on y (some measure of behavior or consciousness)? Testing comes from another tradition, which only became clarified by Alfred Binet, and afterwards: How do people differ in performance on the same task; and what do those differences mean?
2. Galton, Individual differences and Mental tests. He studied reaction times and sensory acuity, which he believed correlated with powerful and efficient nervous systems. He also studied individual differences in memory span and free associations. Galton discovered the concepts of correlation and regression, the two major statistical ideas in test construction and evaluation. His student Karl Pearson developed the Pearson Product Moment Correlation Coefficient (r).
3. Alfred Binet (1857-1911) is most important in developing the first intelligence tests. He was commissioned by the Parisian school system to help identify students who might not profit from regular schooling. Rather than use theoretical ideas of what intelligence is, he used empirical measures with criterion groups. He found out from teachers which students were deficient, and tried to find tests that would differentiate them from normal students. Discrimination and reaction time measures did not work well. He got mean differences between normal and deficient with some measures of a) attention, b) memory, c) following directions, d) making up sentences, e) general knowledge, f) finding logical absurdities, g) reasoning, and others. After finding out that no questions generally discriminate the deficient from the normal child, Binet got the idea that the tests could be age related. Older children were able to answer harder questions than younger. Also, deficient students were likely to be able to answer the same questions as younger normal students. He developed a test that was graded in difficulty, which he scored by whether the average child of a given age could answer it. He identified a child’s intellectual level by comparing her performance with that of the average child. If a child could not generally answer questions that a child two years younger could s/he was deemed retarded, and in need of special instruction. Binet thought that there are many different reasons that a child might not perform, such as illness, a different culture, and poor environment, and stressed that performance of equivalently intelligent children could be patterned very differently.
4. William Stern introduced the term of intelligence quotient by dividing mental age by chronological age.
5. Charles Spearman (1863-1945) found tests to be positively (and hierarchically) correlated with one another. Using a statistical method he helped develop called "factor analysis" he introduced the concept of g or "general intelligence" which represented the capacity for intellectual work, and several specific (s) factors for "specific neurological engines."
6. Lewis Terman (1877-1956) introduced the abbreviation IQ and multiplied the quotient by 100 to rid it of a decimal point. This defines the Average IQ as 100. Terman also revised the Binet test, and standardized it on American children (the Stanford-Binet) so that boys and girls would have equal average IQ. Terman was interested in high rather than low scorers. He estimated the IQ’s of famous people and gave the Stanford-Binet to many children in California. He followed 1500 children who scored 140 or better (they are still being studied) and found they generally, but not entirely, succeeded quite well.
7. Problems with unidimensional view of intelligence (some sub-tests don’t correlate very well with other tests), prodigies in particular areas, specific deficits from accidents. Relatively weak correlations between IQ and certain other measures.
8. Almost from the beginning there have been problems with MA/CA formula, such as significant variations in variance and ceiling effects. David Wechsler (1896-1981) developed IQ tests that measured deviation IQ’s instead of ratio ones. This is important since measured "mental age" does not increase after the mid-teens. By the 1960’s IQ in all tests has been based on deviation scores rather than ratios, although Binet’s concept of graded difficulty still is used.
9. Spearman, Stern, Terman, Goddard, Burt, Yerkes and others supported a strongly hereditarian view of IQ. Many people in clinical and social psychology, as well as anthropologists and others opposed them. Stephen J. Gould and Leon Kamin evaluated much of the research supporting the hereditary view in the 1960's and found it came up short. They found many biases in the presentations of the data, including not only biased selection of data, but in some instances, outright fraud such as the actual manufacturing of data. However, recent data does seem to show that if no drastic social or physiological events occur, IQ measures over time correlate to a great extent with genetic overlap, and only marginally with environmental influences.
10. Henry Goddard (1866-1957) and others used data from IQ tests given on Ellis Island to immigrants to bolster their view that Northern Europeans are genetically the smartest peoples there are. They reported that up to 70% of the Eastern European immigrants were feebleminded. Herrnstein and Murray in The Bell Curve use their beliefs in the innateness of IQ to argue for conservative social policy.
11. Louis Thurstone (1887-1955) using a different variant of Factor Analysis from that of Spearman analyzed correlations from intelligence test items and argued that there was no clear single measure of intelligence. Rather, he found seven "primary mental abilities" which were relatively independent of one another.
12. Recently Howard Gardner has argued that one cannot use a single test to identify intelligence. By considering a set of principles including functionality, relative independence, some symbolic base, some neurological correlates, examples of prodigies, some examples of geniuses and perhaps others, Gardner has been investigating some seven "intelligences" including linguistic, logico-mathematical, spatial, musical, kinesthetic, internal social, and external social. Such individuals as Freud, Larry Bird, Martha Graham, Mozart, Gandhi, represent different kinds of geniuses.
13. Robert Sternberg is another researcher who argues for diverse intelligences. He finds a different set from those of Gardner. Sternberg questions what is meant by intelligence and thinks that Gardner may misuse the term. Neither Sternberg nor Gardner believes that IQ tests identify intelligence.

Other uses of tests and correlation.

1. The ideas of reliability and validity led to many uses for tests. Tests and testing are exceedingly important in education and clinical psychology. Industrial Engineering and Industrial Psychology use tests and other measures and statistical methods for selection of people and evaluation of products and processes.
2. Paul Meehl’s Statistical vs. Clinical Prediction. Lack of strong subjective predicting power. If data correlate significantly with a criterion, they almost always serve as a better predictor than any form of subjective judgment based on interactions with an individual.
3. Applications of psychology in other arenas are still developing. It is used in design of many products from car seats and dashboards to digital phones and computer recognition.
4. This instructor thinks that the use of the normal curve as a basis of a single factor of intelligence is noteworthy. The brain is the basis of intelligence. There are over a billion genes that play a role in the brain’s development. A normal curve is based on a large number of independent factors.