2.8 Intelligence and Achievement

There's no single agreed definition. One view, starting with Charles Spearman, holds that a general intelligence factor, called g, underlies performance on many different tasks, since people who do well on one kind of mental test tend to do well on others. Other theorists argue intelligence is made of several distinct abilities. Howard Gardner proposed multiple intelligences, such as linguistic, logical-mathematical, musical and interpersonal. Robert Sternberg proposed three: analytical, creative and practical.

Because definitions are debated, and because the people who design tests bring their own cultural assumptions, measuring intelligence can be biased.

Early tests, starting with Alfred Binet's work in France in the early 1900s, estimated a child's mental age: the age level at which they performed. The original intelligence quotient (IQ) was mental age divided by chronological age, times 100. A 10-year-old performing like a typical 12-year-old scored 12 ÷ 10 × 100 = 120. This formula breaks down for adults, so modern tests compare you with others your age instead. Scores are set so the average is 100, with most people falling between 85 and 115. Schools today often use IQ scores as one piece of information when deciding which students qualify for special educational services.

A useful test must follow sound psychometric principles (the science of measuring mental traits):

• Standardized: given and scored with the same procedures and conditions for everyone, and compared with a large, representative group that took the test earlier.
• Reliable: gives consistent results. Test-retest reliability means similar scores when the same person retakes the test. Split-half reliability means scores on two halves of the test (like odd versus even questions) agree.
• Valid: measures what it claims to measure. Construct validity means it really measures the trait (the 'construct') it's meant to. Predictive validity means scores predict the future performance they're supposed to, like grades.

Stereotype threat is the worry that you might confirm a negative stereotype about your group, which can lower performance. Stereotype lift is a boost in performance when people are aware of a negative stereotype about another group they're compared with. Researchers try to build socioculturally responsive tests that reduce both.

The Flynn effect is the rise in average IQ scores across much of the world over the 20th century. The rise was too fast to be genetic, so it points to social factors such as better nutrition, health care, education and living standards.

IQ scores vary more within any group than between groups, so a group average tells you very little about any individual. Poverty, discrimination and educational inequities can lower scores for individuals and groups, and personal and cultural biases can affect how scores are interpreted.

Intelligence scores have been used to limit access to jobs, military ranks and schools, and in the early 1900s were used to argue for restricting immigration to the US, often based on tests that favored people who were fluent in English and familiar with American culture. This connects to the eugenics movement (1.1).

Achievement tests measure what you've already learned, like an end-of-unit exam. Aptitude tests predict how well you'll do in the future, like a test used for admission to a program.

Carol Dweck's research shows that mindset matters. People with a fixed mindset believe intelligence is set at birth, so they may avoid challenges. People with a growth mindset believe ability grows with effort and good strategies, so they tend to persist after setbacks. You won't be tested on labels for specific cognitive abilities or disabilities.