What Is the g Factor? General Intelligence, Spearman, and the Positive Manifold

What Is the g Factor?

If you sit down and take a battery of different mental tests — one on vocabulary, one on pattern completion, one on mental arithmetic, one on spatial rotation — something curious happens. People who score well on one of these tasks tend to score well on the others too, even though the tasks look nothing alike on the surface. This tendency for performance across very different cognitive tasks to move together is the empirical starting point for one of the most durable ideas in psychology: the g factor, short for general intelligence.

The g factor is not a single "thing" you can point to in the brain, like a muscle or an organ. It is a statistical concept — a common thread that factor analysis extracts when you look at how scores on many different cognitive tests relate to one another. In practice, most standardized IQ tests are built from several subtests (verbal reasoning, working memory, processing speed, visual-spatial reasoning, and so on), and the overall IQ score is essentially an estimate of this shared, general factor running through all of them.

The Positive Manifold: Why Different Tests Correlate

The observation that almost all cognitive tests correlate positively with one another — rarely negatively, rarely at zero — is known as the positive manifold. It is one of the most consistently replicated patterns in psychological measurement. Someone who is unusually good at solving abstract number sequences is, on average, also somewhat better than chance at verbal analogies, spatial puzzles, and working-memory tasks, even though these skills draw on very different content.

This does not mean every person is equally good at everything, and it does not mean specific abilities do not exist. A person can have a real strength in, say, spatial reasoning relative to their own verbal reasoning. What the positive manifold shows is that beneath those individual differences in specific skills, there is a broad, general tendency that pulls performance across tasks in the same direction. That general tendency is what statisticians label g.

Spearman's Two-Factor Theory

The British psychologist Charles Spearman is credited with formalizing this idea in the early twentieth century, when he noticed the pattern of positive correlations among school children's performance across different academic and sensory-discrimination tasks. He proposed what became known as the two-factor theory of intelligence: every cognitive test score reflects a combination of (1) the general factor, g, which contributes to performance on virtually all cognitive tasks, and (2) a specific factor, s, unique to that particular test or skill.

Spearman's method — factor analysis — was itself a methodological breakthrough. It gave psychologists a mathematical tool to ask whether a bundle of correlated test scores could be explained by one or a few underlying dimensions rather than treating every test as measuring something entirely separate. Later researchers built on and revised this model considerably, proposing hierarchies with broader group factors (like verbal comprehension or processing speed) sitting between specific test scores and g at the top. But the core insight — that a general factor accounts for a substantial share of the shared variance across cognitive tests — has proven remarkably robust across the decades and across different test batteries.

Fluid and Crystallized Intelligence

One influential way of unpacking g further distinguishes between fluid intelligence and crystallized intelligence. Fluid intelligence refers to the capacity to reason and solve novel problems independent of previously learned knowledge — pattern recognition, abstract reasoning, and logical inference are classic examples. Crystallized intelligence, by contrast, reflects accumulated knowledge and skills built up through education, culture, and experience, such as vocabulary size or factual knowledge.

These two broad abilities are themselves positively correlated (consistent with the positive manifold) and are often treated as major components that feed into the overall general factor. Many modern IQ tests, including the reasoning-style items you'll find on IQTesta, lean heavily on fluid-intelligence tasks such as pattern completion and matrix reasoning, partly because these tasks are less dependent on specific language or cultural background than crystallized-knowledge questions.

Why This Matters — And Its Limits

Scores from tests designed to estimate g are typically scaled so that the general population has an average score of 100 with a standard deviation of 15, following a roughly normal (bell-curve) distribution. This scaling is why a score of 130 or higher — associated with organizations like Mensa, whose membership threshold sits around the 98th percentile — is understood as substantially above average, while a score of 70 or below is substantially below it.

It's also worth knowing that population-level average scores on standardized intelligence tests have shifted over generations in many countries, a well-documented pattern known as the Flynn effect, which is one reason tests are periodically renormed.

Understanding the g factor helps explain why a single IQ score can be a meaningful summary of general cognitive performance rather than an arbitrary number — but it also underscores the limits of any single test, including a quick online one. An IQTesta result is indicative, not a clinical assessment: it can give you a rough, self-informative estimate of general reasoning performance, but it is not a substitute for a full psychometric evaluation administered by a qualified professional. This caution matters especially when the subject is a child or when the question involves giftedness or a possible learning difficulty — in those cases, any concerns should be discussed with a psychologist or other qualified specialist rather than settled by an online test.