The Nature of Intelligence
In the last module we looked at the history of intelligence testing. What the tests actually measure is the subject of this module. Some of the important factors that make up what we call intelligence can be statistically identified. These include verbal and numerical skills, memory, reasoning ability, spatial relations, and so on.
As you read the text, try to answer the following questions:
One Factor or Many?
Most of us think of intelligence as an overall ability or attribute of a person. That is, we tend to classify people as either bright, average or dull. It is much less common to hear someone differentiate between conditions of intelligence. Of course, anyone who wishes to develop an intelligence test must come to grips with this problem. Is there such a thing as general intelligence as distinct from specific abilities, or is intelligence just a collection of abilities? Binet saw intelligence as one basic quality
Binet assumed that intelligence was essentially unitary; that is, he maintained that the mind possesses one overriding quality, which he believed to beability to effectively adjust to the environment. In devising his test he assumed that this unitary factor could be manifested in three ways. (1 ) As the ability to maintain a definite direction. (2) In choosing means appropriate to ends. (3) In objectively evaluating one's own actions.
Charles Spearman, an English statistician, looked at intelligence differently. He devised a two-factor theory. On the basis of his correlational studies he concluded that intelligence could be divided into a general factor (G) and a group of specific factors (S1, S2, S3, etc.). His theory was that "G" played a part in all intellectual activities, but also that any particular activity was affected by one or more of the specific factors.
E. L. Thorndike rejected the notion that there was such a thing as a general intelligence factor. He formulated the multifactor theory. This theory stated that intelligence is made up of many independent components, all of which, when added together, form intelligence. Thorndike took the position that a person could be very strong in one factor and very weak in another, and be just as intelligent as someone else who had the reverse capabilities.
A fourth theory, proposed by L. L. Thurstone, which is somewhere in between those just presented. According to his theory, intelligence is not just a matter of "G" plus "S" but a primary group of factors (6 or 7) and one general or "G" factor. Thurstone arrived at this position as a result of his factor analysis studies of intelligence. .
Thurstone's work indicated that a number of factors go into making up intelligence and further that each of the factors are In part related to each other and to some general intelligence factor (G). .
The multi-factor theories are more popular now, although people disagree as to what the factors are
Factor analysis is a statistical technique designed to identify and isolate the elements that make up any complex trait. A psychologist involved in factor analysis gives different tests to the same people and calculates correlations between responses on subtests or individual items. When all the possible correlations are ready, procedures of statistical analysis are used to find out which items seem to be in clusters and which seem relatively independent. For example, several subtests or items dealing with words may correlate highly with each other, but not at all with subtests or items concerned with numbers. In this example, the statistics indicate that a verbal factor and a numerical factor are relatively independent.
Thurstone conducted an extensive study of the factors of intelligence (Thurstone and Thurstone, 1941). This study, which involved factor analyses of dozens of tests administered to schoolchildren, identified seven intelligence factors:
After deciding that these seven factors made up intelligence, Thurstone rearranged the existing subtests and devised some new ones. He referred to these sets as: "Tests of Primary Mental Abilities," which he used to investigate the question as to whether there was such a thing as "general" intelligence. His Tests of Primary Mental Abilities were administered to a large group of children, and correlations were computed between scores on the various tests. If there were no such thing as general intelligence, each factor would be independent. However, there were positive correlations among all the tests, with some more highly correlated than others, which indicated that primary mental abilities share some common factor. Thurstone concluded that each primary factor is composed of an Independent primary factor and a general factor (G) that is shared by all of the primary factors; in other words, intelligence consists of both general ability and a number of specific abilities.
Today, many psychologists prefer theories by Sternbeg and by Gardner. Both of these psychologists reject the emphasis on a general intelligence. Robert Sternberg (1985) put forward a triarchic theory, proposing that intelligence embodies a wide variety of skills important for efective living. His observation of Yale graduate students, yeilded what he believed were components of intelligence: componential ntelligence (sovling problems; experiential intelligence (adapting creatively); and contextual intelligence (making the environment fit your strengths).
Gardner's theory (1993) is similar to Thurstone's theory. Gardner believes that there are seven separate intelligences: logical-mathematical, linquistic, spatial, musical, kinesthetic, interpersonal, and intrapersonal intelligences.
Later, a group of investigators (Guilford and Hoepfner, 1966) devised some finely differentiated tests that indicated the existence of at least 80, and perhaps as many as 120, separate factors that should be considered elements of intelligence.
The Nature of Adult Intelligence
As we saw in the last module, test scores purporting to define "mental age" increase up to the middle teens. But what happens then? After late adolescence, it appears that intelligence no longer improves and in fact may even deteriorate. This trend has been observed in several studies.
Some studies have shown declining intelligence as people grow older
An early cross-sectional study (Jones and Conrad, 1933) included almost all the inhabitants from age 10 to age 60 in a New England community. The average scores on the First World War army test rose rapidly until about age 18, and then declined steadily from 18 to 60. The scores on separate factors on the test, however, did not all show this same pattern. The scores on the reasoning test showed a sharp decline from age 18 to 30. Thereafter it continued to decline but at a much slower rate. The vocabulary test, however, showed no decline as a factor of age. The rise was rapid until age 20 and then continued to rise very slowly until it reached its highest point at age 60.
Another cross-sectional study that showed a decline in overall IQ scores took place during standardization of the WAIS The peaks were different but the overall picture was similar. That is, the peak of verbal intelligence came in the 25-34 age range and declined more rapidly after age 44 (Wechsler, 1955). These data support the hypothesis that we become less intelligent as we grow older but some people have questioned this conclusion. They argue that the tests are biased against older people, particularly performance tests that place strong emphasis on speed.
Other studies revealed increased intelligence except in skills older people have little reason to use.
Other data lends support to the hypothesis that perhaps the apparent decline in intelligence with age is an artifact of the tests used to measure it. Owens (1953) conducted a longitudinal study with a group of men who had taken the army's group intelligence test during the First World War and then took the same test again thirty years later. Their scores were higher on everything but arithmetic.
One interpretation of all these findings is that each generation grows up in an environment that is more conducive to intellectual growth than the generation before. This would be consistent with the declines observed in cross-sectional studies and the improvements in the longitudinal studies. More longitudinal studies are needed for a better understanding of the relationship between intelligence and age.
Validity of Intelligence Tests
IQ tests are reliable predictors of school success
Usually, we want an intelligence test to have predictive validity. That is, it would be desirable for a test of intelligence to predict how well someone will do on a task that involves intellectual ability. The most frequent use of IQ tests is to predict school success and, in this application, IQ tests generally have high predictive validity. In fact intelligence tests are such good predictors that teachers often become concerned if there is a discrepancy between IQ scores and school success. Students who score high on IQ tests, but who are not doing well in school, are often called "underachievers," reflecting the difference between actual achievement and potential achievement as revealed in the IO test score
However, we should remember that validity is defined as the answer to the question, "Does the test measure what it claims to measure?" In part, that question depends upon how the person constructing the test defines intelligence. For example, if one sees intelligence as the ability to do conceptual thinking, then those test items which require a student to formulate ideas and verbal symbols would be the most appropriate kinds of items to measure this ability. Thus we can talk of a test as having logical validity, which means only that the items we have selected fit our definition of intelligence.
Now test yourself without looking back.
1. A statistical method used to isolate the elements, or factors, that make up intelligence is______________
2. What did Thurstone call the seven separate factors that he found to make up intelligence?______________________
3. At the end of his studies, Thurstone concluded that intelligence was made up of:
a. seven general abilities.
b. seven primary mental abilities
c. a general factor.
4. Up to the middle teens or so, intelligence seems to_________________________________
5. Total intelligence has been found to decline after age 18 in:
a. longitudinal studies.
b. correlation studies.
c. cross-sectional studies.
6. List five of Thurstone's primary mental abilities.___________________________________________
7. If an intelligence test predicts how well one will do in school with a great deal of accuracy, we can say that this test has____________________________________/
In factor analysis we administer intelligence tests and correlate
the results on separate items or subtests. If the results on two
subtests show high correlation, the researcher decides they are:
a. not relevant factors.
b. not independent.
A statistical method used to isolate the elements of intelligence is called________________________________________________________________1
Thurstone found seven primary mental abilities. Write the name of
the ability after each definition.
a. The ability to define and understand words___________________________
b. The ability to produce words rapidly__________________________
c. The ability to solve arithmetic problems______________________
d. The ability to visualize physical relationships___________________________
e. The ability to memorize and recall________________________.
f. The ability to see differences and similarities among objects___________________________
g. The ability to find rules__________________________________
What did Thurstone call the seven basic factors?
Thurstone devised subtests based on his primary mental abilities. He found positive correlation among the subtests. He interpreted this to mean that:
a. there was another primary mental ability.
b. there was a general factor, common to all the primary abilities.
1 factor analysis
2 primary mental abilities
3 b, c
5 a. verbal comprehension
b. word fluency
The relation of mental ability to age in a New England community (After dones and Conrad, 1933)
Based on the data above, complete the following:
a. Total intelligence increases until about age________________________
b. One kind of intelligence that continues to increase is________________________
c. A kind of intelligence that reaches a peak, then drops sharply is__________________________
d. Which method was used in this study?__________________________
A longitudinal study showed no decline in overall intelligence,
although several cross-sectional studies show a pattern of
decline with age. Which of the following could explain this?
a. One type of study is not valid.
b. Young people today might have been reared in a better environment.
c. Bright people die young.
A test has predictive validity if it allows one to predict how a subject will do on a task that involves the use of the capabilities that the test measures. In the text we pointed out that IQ tests are generally good predictor of scholastic success. Suppose we had developed four different welders' aptitude tests in order to predict future success in the welding profession. We test our subjects and then correlate these scores with the ratings made by their supervisors as to the quality of the work. Which test has the highest predictive validity?
a. Test A correlates .04 with supervisor's rating.
b. Test B correlates -.71 with the supervisor's rating.
c. Test C correlates .52 with the supervisor's rating.
d. Test D correlates .28 with the supervisor's rating.
NOW TAKE PROGRESS CHECK 2
4. a. 18
1. Factor analysis is used to:__________________________________________________
2. Thurstone's studies are noted because they:
a. use factor analysis
b. provide proof of the decline in intelligence after age 18.
c. identify several general factors in intelligence.
d. (none of these)
3. What did Thurstone call the factors that make up intelligence__________________
4. Name five of Thurstone's primary mental abilities._______________________________________________________________________________
5. General intelligence has not been shown to decrease with age in _______________________
6. What might be a reason for a decline in measured intelligence with age in cross-sectional studies?
7. An intelligence test has predictive validity if it:
a. appears to measure the type of ability its designers claim it should.
b. predicts the success of individuals on a task that requires intellectual ability.
c. predicts who will make a similar score on a retest.
d. predicts how the designer views intelligence.
Table of Contents for Unit 7
March 8, 2002