IQ scores rose dramatically throughout the twentieth century, climbing roughly three points per decade in many industrialized countries. That trend, known as the Flynn effect, made each generation score higher than the last on standardized intelligence tests. But starting around the mid-1990s, researchers began detecting slowdowns and even reversals in a handful of nations, sparking headlines about humanity getting dumber. The reality is more interesting and more unsettling than a simple yes or no: the forces shaping measured intelligence are overwhelmingly environmental, which means the answer depends less on our biology than on the conditions we create for our brains.
A Century of Rising Scores
The Flynn effect is one of the most robust findings in psychology. A meta-analysis spanning 285 studies and more than 14,000 participants found that IQ scores climbed an average of about 2.3 standard-score points per decade, with modern Wechsler and Stanford-Binet tests showing gains closer to three points per decade.1PubMed Central. The Flynn effect: a meta-analysis A separate meta-analysis covering data from 31 countries and nearly four million participants across more than a century (1909–2013) confirmed that gains were worldwide, though they varied by the type of ability being tested. Fluid reasoning, the kind of abstract problem-solving that doesn’t depend on learned knowledge, gained the most, at roughly 0.41 IQ points per year. Crystallized intelligence, which reflects accumulated facts and vocabulary, gained the least.2PubMed. One Century of Global IQ Gains: A Formal Meta-Analysis of the Flynn Effect (1909-2013)
What drove those gains? Researchers have pointed to the massive expansion of formal schooling, improvements in public health, smaller family sizes, and a more cognitively demanding modern environment. The education hypothesis is especially compelling because schooling trains exactly the kind of abstract thinking that IQ tests reward: planning, spatial reasoning, working memory, and unique problem-solving.3Intelligence. The cognitive impact of the education revolution: A possible cause of the Flynn Effect on population IQ One thing researchers have broadly agreed on is that whatever produced these gains, it was not a change in the gene pool. Evolution works far too slowly to explain score jumps that happened within a single generation.
Where Scores Have Started Falling
The first signs that the Flynn effect might be stalling came from Scandinavia. A systematic review using strict quality criteria found negative Flynn effects reported in seven countries, meaning average test scores were going down rather than up.4Intelligence. The negative Flynn Effect: A systematic literature review Norway’s military conscription data, which covers virtually every young man in the country, offered an unusually detailed picture. Scores rose for cohorts born in the early 1960s, peaked for those born around the mid-1970s, and then declined for later birth cohorts.5PubMed Central. Flynn effect and its reversal are both environmentally caused The broader meta-analysis of worldwide data also noted that gains have decreased in more recent decades, even though they haven’t disappeared everywhere.2PubMed. One Century of Global IQ Gains: A Formal Meta-Analysis of the Flynn Effect (1909-2013)
It is worth being specific about the geography here. The reversal has been most clearly documented in wealthy Northern European countries with excellent conscription data. Many lower- and middle-income nations are still experiencing rising scores as access to education, nutrition, and healthcare continues to expand. The picture is not one of uniform global decline.
Why the Changes Are Environmental, Not Genetic
A common fear embedded in this question is eugenic in flavor: are less intelligent people outbreeding more intelligent ones, dragging down the species? The evidence strongly argues against that explanation. The Norwegian conscription study demonstrated that the entire Flynn effect, its peak, and its subsequent decline could be recovered from variation within families, meaning brothers born to the same parents and raised in the same household scored differently depending on when they were born.5PubMed Central. Flynn effect and its reversal are both environmentally caused If the changes were genetic, you would not see them between siblings sharing the same parents.
A Dutch study of military brothers reinforced this. When average IQ scores were rising nationally, later-born brothers scored higher than their older siblings. When scores were declining, later-born brothers scored lower. And in periods with no national trend, birth order didn’t predict score differences.6Intelligence. The Flynn effect in sibships: Investigating the role of age differences between siblings Whatever is pushing scores around is something in the shared environment that shifts from year to year, not something inherited at conception.
Lead Exposure and IQ at the Population Level
If you want a concrete example of how the environment shapes intelligence on a mass scale, lead is the most sobering case study. An international pooled analysis of children from seven countries found that an increase in blood lead from about 2 to 10 micrograms per deciliter was associated with a loss of nearly four IQ points. Crucially, the damage was proportionally steepest at the lowest exposure levels, meaning that even small amounts of lead do outsized harm to a developing brain.7PubMed Central. Low-level environmental lead exposure and children’s intellectual function: an international pooled analysis An earlier meta-analysis found that raising blood lead from 10 to 20 micrograms per deciliter was associated with a drop of about 2.6 IQ points and argued the relationship should be considered causal, given that lead disrupts neurotransmitter systems involved in learning and memory.8PubMed. Low-level lead exposure and children’s IQ: a meta-analysis and search for a threshold
The scale of lead’s impact in the United States alone is staggering. Researchers estimated that, as of 2015, childhood lead exposure was responsible for the cumulative loss of over 824 million IQ points across the U.S. population, with those born in the leaded-gasoline era hit hardest.9PubMed Central. Half of US population exposed to adverse lead levels in early childhood The phase-out of leaded gasoline and lead paint, which accelerated in the 1970s and 1980s, likely contributed to the Flynn effect’s continuation during that period by removing a major drag on childhood brain development. But lead is not the whole story, and it certainly does not explain the reversals seen in Scandinavian countries where lead exposure was already low.
Indoor Air and Cognitive Performance
A less obvious environmental factor getting increased attention is the air inside buildings. Most people spend the vast majority of their time indoors, and indoor carbon dioxide levels in poorly ventilated classrooms and offices can climb well above outdoor concentrations. A controlled exposure study found that at roughly 1,000 parts per million of COâ‚‚, modest but statistically significant decrements appeared across six of nine decision-making scales compared to 600 ppm. At 2,500 ppm, large reductions appeared on seven scales.10PubMed Central. Is CO2 an indoor pollutant? Direct effects of low-to-moderate CO2 concentrations on human decision-making performance Another controlled study of office workers found that cognitive function scores dropped about 15% at around 945 ppm and about 50% at around 1,400 ppm compared to well-ventilated conditions.11PubMed Central. Associations of Cognitive Function Scores with Carbon Dioxide, Ventilation, and Volatile Organic Compound Exposures in Office Workers: A Controlled Exposure Study of Green and Conventional Office Environments
A systematic review and meta-analysis looking across multiple studies confirmed that COâ‚‚ exposure below 5,000 ppm impairs cognitive performance, with complex tasks hit harder than simple ones. Children in schools with elevated COâ‚‚ also showed decreases in task accuracy.12Building and Environment. Short-term exposure to indoor carbon dioxide and cognitive task performance: A systematic review and meta-analysis These are short-term effects on performance rather than permanent changes to intelligence, but they matter for a practical reason: if students spend years learning in stuffy classrooms, the cumulative effect on skill acquisition could plausibly show up in test scores. Rising outdoor COâ‚‚ levels from fossil fuel combustion will continue pushing indoor concentrations higher unless ventilation standards keep pace, making this a slow-moving problem worth watching.
What Technology Is and Isn’t Doing to Our Brains
It is tempting to blame smartphones for making everyone dumber, and there is a kernel of truth here, but it is smaller and stranger than the popular narrative suggests. A study of Brazilian university students found that problematic smartphone use was correlated with self-reported complaints about memory, attention, and self-control. However, when those same students took objective tests of working memory and inhibitory control, no consistent impairments showed up.13Journal of Technology in Behavioral Science. The Relationship Between Problematic Smartphone Use and Difficulties in Attention, Working Memory, and Inhibitory Control Among Brazilian University Students In other words, heavy phone users felt scatterbrained but did not actually perform worse on lab tasks. That gap between subjective complaint and objective performance is a recurring finding that should give pause to anyone making sweeping claims about phones melting brains.
The more concerning finding involves what researchers call cognitive offloading: the habit of saving information externally rather than committing it to memory. A study of both children and adults found that when participants expected to have access to a saved list, they recalled fewer words and reported using fewer memory strategies. Children aged 10-11 showed the same pattern as adults, reducing their internal encoding when they could rely on an external aid.14PubMed Central. Cognitive offloading reduces internal memory processing in children This does not mean Google is destroying memory capacity. It means our brains are doing what they have always done: allocating effort efficiently. If you know something is saved, you don’t bother memorizing it. The practical risk is that children who constantly offload to devices may not develop the kind of deep encoding strategies that help with learning in domains where external aids aren’t available, like understanding a concept rather than recalling a fact.
Pandemic School Closures and Measurable Learning Loss
The COVID-19 pandemic created a natural experiment in what happens when you abruptly remove formal education. Global data from the PISA assessment showed that mathematics scores declined by an average of about 14 percent of a standard deviation, equivalent to roughly seven months of learning. The losses were worse for students in schools with longer closures, and disproportionately affected boys, immigrants, and disadvantaged students.15Nature Partner Journals / PubMed Central. COVID-19, school closures, and student learning outcomes. New global evidence from PISA This finding serves as a real-time demonstration of the relationship between educational environment and cognitive performance. Whether those losses will be recovered as schooling returns to normal, or whether a generation of students will carry a permanent deficit, is one of the most consequential open questions in education research right now.
Are the Tests Themselves the Problem?
Here is where the story gets genuinely uncomfortable for anyone trying to draw firm conclusions. The Flynn effect has always been measured by comparing scores on IQ tests normed at different times. But that approach assumes the tests are measuring the same underlying abilities in the same way across decades, and that assumption does not always hold up. A study examining three editions of the Wechsler Adult Intelligence Scale found that while scores were comparable between the third and fourth editions, they were not equivalent between the second and third editions. The researchers concluded that there was some evidence for a real increase in intelligence but cautioned that many published Flynn effect findings rest on the unwarranted assumption that test scores from different eras are directly comparable.16PubMed. Using Score Equating and Measurement Invariance to Examine the Flynn Effect in the Wechsler Adult Intelligence Scale
A recent analysis of Norwegian military data pushed this further. When researchers applied formal tests of measurement invariance, they found that neither the original Flynn effect nor its reversal met the statistical criteria for scalar invariance, meaning the tests might not have been measuring the same construct in the same way across time periods.17Intelligence. Reevaluating the Flynn effect, and the reversal: Temporal trends and measurement invariance in Norwegian armed forces intelligence scores If the yardstick itself has been changing, then some portion of both the rise and the decline in scores may be an artifact of the measurement rather than a reflection of genuine cognitive change. This does not mean the Flynn effect is fake; it means the size of the effect, and especially the size of its reversal, may be less certain than the raw numbers imply.
Sleep Schedules and Everyday Cognitive Function
Beyond the big population-level trends, individual cognitive performance fluctuates based on something most people ignore: how well their sleep schedule lines up with their biological clock. Social jet lag, the mismatch between when your body wants to sleep and when your social obligations (work, school, alarm clocks) force you awake, acts as a form of chronic mild sleep deprivation. A study using real-world cognitive testing found that social jet lag was consistently the strongest predictor of cognitive performance among the factors examined, and the impairment was stable across days. It also increased the rate of false positives on attention tasks, suggesting reduced ability to suppress automatic responses.18PubMed. Sleep Misalignment and Cognitive Decline in Everyday Life-Social Jet Lag as a Proxy for Chronic Sleep Deprivation
This matters for the broader intelligence question because modern life systematically pushes people toward sleep misalignment. Evening screen use, irregular work schedules, and early school start times all widen the gap between biological and social time. If a large fraction of the population is chronically slightly sleep-deprived, their cognitive test scores will understate their actual capability, just as poor indoor air quality drags down performance without permanently altering the brain.
Nutrition Is Not the Simple Answer It Seems
Better nutrition is often cited alongside education as a driver of the Flynn effect, in part because average height also increased over the twentieth century and height gains are clearly linked to improved diet. But the two trends may not be as connected as that parallel suggests. An analysis of British Raven’s Progressive Matrices scores from 1938 to 2008 found that IQ gains and height gains did not track each other closely, and the specific pattern of IQ gains could not be accounted for by nutritional improvements alone.19PubMed. Requiem for nutrition as the cause of IQ gains: Raven’s gains in Britain 1938-2008 Severe malnutrition undeniably harms brain development, but beyond a threshold of adequate nutrition, adding more calories or micronutrients does not appear to keep boosting cognitive scores. This is one reason why the Flynn effect has plateaued in wealthy countries where nutrition was already excellent by the mid-twentieth century.
The Shrinking Human Brain
If you zoom out from decades to millennia, there is one genuinely startling fact. Human brains have gotten physically smaller. Since the Late Pleistocene, roughly 30,000 years ago, average brain volume has decreased by about 10 percent, although body size decreased in parallel, so the brain-to-body ratio did not change as dramatically.20PubMed. Evolution of the human brain: is bigger better? This is sometimes cited as evidence that modern humans are less intelligent than their ancestors, but that interpretation is a stretch. Brain size is a crude proxy for cognitive ability. The relationship between raw brain volume and intelligence within a species is weak, and brains can become more efficient through changes in neural connectivity, cortical folding, or the proportion of different tissue types. A smaller engine that runs more efficiently can still outperform a bigger one.
It is also possible that the reduction in brain size reflects the fact that survival in complex societies requires different cognitive skills than survival as a Pleistocene hunter-gatherer. Cooperation, language, and cultural transmission allow modern humans to distribute cognitive labor across a group rather than relying on any single individual’s raw processing power. The brain may have become slightly smaller precisely because cultural infrastructure took over some of the work that raw neural tissue used to perform.
Distinguishing Real Decline from Shifting Demands
Much of the anxiety around this question conflates different things. A dip in standardized test scores is not the same as a decline in human cognitive capacity, just as a drop in your step count doesn’t mean your legs stopped working. Many of the factors associated with score declines, such as indoor air quality, sleep misalignment, screen-based distraction, and pandemic-related education disruption, are performance impairments rather than permanent losses. Remove the cause and, in principle, performance recovers.
There is also a persistent confusion between intelligence and particular cognitive skills. If young adults today are worse at mental arithmetic than their grandparents were but better at filtering information from multiple sources and navigating complex digital interfaces, has intelligence declined or changed shape? IQ tests necessarily measure a specific set of abilities at a specific moment. They do a good job of predicting academic and occupational outcomes within a given era, but they were never designed to be timeless barometers of human potential across generations. The measurement invariance problems found in both the Wechsler and Norwegian military data remind us that comparing scores across decades is harder than it looks, and that some amount of the apparent decline may reflect shifting test demands rather than shifting brains.
Where there are genuine grounds for concern is in the environmental exposures that damage developing brains irreversibly. Childhood lead exposure does not just temporarily depress test performance; it causes lasting structural harm. To the extent that populations were heavily exposed to lead (or other neurotoxicants) during key developmental windows, the IQ loss is real and not recoverable. The United States is still living with the cognitive toll of decades of leaded gasoline, and many lower-income countries face ongoing exposure from informal recycling, contaminated water, and legacy industrial pollution. For those populations, the question is not whether humans are becoming less intelligent in some abstract sense. The question is whether avoidable environmental harm is robbing millions of children of cognitive potential they would otherwise have had.