Because the psychometric approach focuses on the interrelationships among intellectual abilities, the major goal is to describe the ways in which these relationships are organized (Sternberg, 1985). This organization of interrelated intellectual abilities is termed the structure of intelligence. The most com­mon way to describe the structure of intelligence is to picture it as a hierarchy (Cunningham, 1987).

Each higher level of this hierarchy represents an attempt to organize components of the level below in a smaller number of groups. The low­est level consists of individual test questions—the specific items that people answer on an intelligence test. These items can be organized into tests at the second level. The third level, primary mental abilities, reflects interrelationships among perfor­mances on intelligence tests. The interrelationships uncovered among the primary mental abilities pro­duce the secondary mental abilities at the fourth level. Third-order mental abilities in turn repre­sent i nterrelationships among the secondary men­tal abilities. Finally, general intelligence at the top refers to the interrelationships among the third – order abilities.

Keep in mind that each time we move up the hierarchy we are moving away from people’s actual performance. Each level above the first represents a theoretical description of how things fit together. Thus there are no tests of primary abilities per se; primary abilities represent theoretical relationships

240 CHAPTER 7 among tests, which in turn represent theoretical relationships among actual performance.

So exactly how do researchers construct this theoretical hierarchy? The structure of intelli­gence is uncovered through sophisticated statis­tical detective work using factor analysis. First, researchers obtain people’s performances on many types of problems. Second, the results are exam­ined to determine whether performance on one type of problem, such as filling in missing letters in a word, predicts performance on another type of problem, such as unscrambling letters to form a word. If the performance on one test is highly related to the performance on another, the abilities measured by the two tests are interrelated and are called a factor.

Most psychometric theorists believe that intel­ligence consists of several factors. However, we should note that although factor analysis is a sophis­ticated statistical technique, it is not an exact tech­nique. Thus estimates of the exact number of factors vary from a few to over 100. Most researchers and theorists believe the number to be relatively small. We will examine two types of factors: primary and secondary mental abilities.

Primary Mental Abilities. Early in the 20th century researchers discovered the existence of several independent intellectual abilities, each indicated by different combinations of intelligence tests (Thurstone, 1938). The abilities identified in this way led to the proposition that intelligence is composed of several independent abilities, labeled primary mental abilities. Thurstone ini­tially examined seven primary mental abilities: number, word fluency, verbal meaning, associa­tive memory, reasoning, spatial orientation, and perceptual speed. Over the years this list has been refined and expanded, resulting in a cur­rent list of 25 primary mental abilities that have been documented across many studies (Ekstrom et al., 1979). Because it is difficult to measure all 25 primary abilities in the same study, research­ers following in Thurstone’s tradition concentrate on measuring only a subset. Typically, this subset originally consisted of the following five primary mental abilities:

• Numerical facility, or the basic skills underlying one’s mathematical reasoning

• Word fluency, or how easily one can produce verbal descriptions of things

• Verbal meaning, or one’s vocabulary ability

• Inductive reasoning, or one’s ability to extrapolate from particular facts to general concepts

• Spatial orientation, or one’s ability to reason in the three-dimensional world in which we live

Two other important information-processing abil­ities, perceptual speed and verbal memory, were incorporated into the battery of measures in subse­quent work (Schaie, 1994, 1996):

• Perceptual speed is one’s ability to rapidly and accurately find visual details and make comparisons.

• Verbal memory refers to the ability to store and recall meaningful language units. Note that tests for this ability typically include word fluency. In these analyses, word fluency is eliminated as a separate ability.

Next we consider how these abilities fare with age.