Reaction Time and IQ: Why Speed of Processing Predicts Intelligence

The Unexpected Predictor

If you wanted to predict someone's IQ without using an IQ test, one of the most reliable things you could measure is their reaction time — specifically, the variability of their reaction times across many simple trials. This is counterintuitive. Pressing a button when a light turns on seems like it has nothing to do with abstract reasoning, pattern recognition, or verbal comprehension. And yet the correlation is real, consistent across studies, and has generated decades of research.

The correlation between simple reaction time (RT) and IQ is approximately r = -0.20 to -0.30 (negative because faster RT = higher IQ). Intraindividual variability in RT — how consistent your reaction time is across many trials — correlates even more strongly with IQ, typically r = -0.30 to -0.45.

The history here is stranger than it sounds. Francis Galton proposed reaction time as a measure of intelligence in the 1880s, before IQ tests existed. He assumed faster neural transmission meant better mental equipment and collected data from nearly 10,000 subjects at his South Kensington lab. His analysis showed essentially no relationship between RT and his other ability measures, and the idea was mostly abandoned for a century. Arthur Jensen revived it in 1982 with properly controlled experiments and modern timing equipment, finding the correlations that Galton's crude methodology couldn't detect. Galton had the right hypothesis. His tools just weren't precise enough.

Why Speed Predicts Intelligence

Several theories attempt to explain the RT–IQ relationship. The two most supported are:

Neural efficiency hypothesis. Higher intelligence may reflect more efficient neural information processing — faster and more consistent signal transmission, lower neural noise, and more precise encoding of information. Reaction time is a direct measure of this processing efficiency, uncontaminated by strategic thinking or domain knowledge. On this view, RT and IQ are both measuring the same underlying neural property from different angles.

Inspection time studies. A related paradigm measures how briefly a visual stimulus needs to be presented for a subject to reliably identify it. "Inspection time" correlates with IQ at approximately r = -0.50 — one of the strongest correlations between a simple perceptual measure and general intelligence ever found. This suggests the relationship runs deeper than motor speed to something about perceptual processing efficiency itself.

The reason intraindividual variability is a stronger predictor than average RT is worth understanding. IIV measures how much your reaction times bounce around across trials. A highly intelligent person doesn't just respond faster on average — they respond more consistently. Researchers interpret this as reflecting neural noise: lower noise means less trial-to-trial variation. Someone with high IQ might average 230ms with a standard deviation of 25ms; someone with lower IQ at the same average speed might show 45ms variability. The consistency does much of the predictive work, which is why a single reaction time measurement tells you almost nothing, but 200 trials across a session tell you quite a bit.

What This Does Not Mean

The RT–IQ relationship does not mean that people with faster reflexes are smarter, or that reaction time training raises IQ. Correlation in population studies does not imply that individual-level variation in RT is the cause of variation in IQ.

Several important caveats apply:

• The correlation is moderate, not strong. Reaction time explains perhaps 10% of variance in IQ across large samples — meaningful statistically, but not diagnostic at the individual level
• Simple RT (one stimulus, one response) shows weaker correlations than choice RT or intraindividual variability measures
• RT is sensitive to fatigue, alertness, and motivation in ways that make it unreliable as an individual assessment tool
• Athletes and trained individuals show faster reaction times without corresponding IQ advantages — motor training and neural efficiency are separable

Processing Speed in IQ Testing

Most comprehensive IQ batteries include a processing speed subtest precisely because of this relationship. The WAIS-IV processing speed index, for example, measures how quickly and accurately a person can process simple visual information. It correlates with the other cognitive factors in the battery and contributes to the full-scale IQ score.

What the RT research ultimately suggests is that general intelligence has a genuine biological substrate — that it is not purely a cultural construct or a product of education — and that some component of what we measure as intelligence reflects basic neural processing efficiency. The practical implications of this are modest, but the theoretical implications for understanding what intelligence actually is are significant.

Reaction Time and Cognitive Aging

Probably the most practically important application of the RT-IQ relationship is in aging research. Processing speed declines reliably with age — simple RT in healthy 70-year-olds is typically 20-40% slower than in healthy 25-year-olds. Timothy Salthouse's processing speed theory of adult cognitive aging argues that this slowdown is the primary engine driving age-related fluid intelligence decline, not deterioration in specific cognitive systems. As neural transmission becomes less efficient, everything downstream — working memory, reasoning, the ability to integrate multiple pieces of information — suffers as a consequence.

This framing has clinical implications. Changes in processing speed measured by RT tasks can detect early cognitive decline before it becomes visible on standard IQ assessments. Some longitudinal studies show that RT variability increases years before other cognitive symptoms appear, which makes it a candidate biomarker for conditions like early-stage dementia. The humble button-press task turns out to be measuring something that matters across a whole lifespan.