Average IQ Score: What the Number 100 Actually Means
The average IQ score is 100 on every major standardised test — not because that is a natural midpoint of human reasoning, but because every test is deliberately calibrated to place its reference population there. Understanding that design choice changes how you should read any score you receive.
This guide covers the full picture: how score ranges map to percentiles, why the average has shifted over generations (the Flynn Effect), what cross-country data actually reveals, and how to interpret your own result against a meaningful reference point.
IQ scores follow a normal distribution within a normed population. About 68% of people score between 85 and 115 (one standard deviation either side of 100), and about 95% fall between 70 and 130. These proportions are fixed by the mathematics of the normal distribution — not by observation of any specific population.
68% (1 SD): 85 – 11595% (2 SD): 70 – 130
What Is the Average IQ Score?
The average IQ score on any well-constructed standardised test is exactly 100. This is not a coincidence or an approximation — it is a deliberate calibration decision. When test designers norm a new instrument, they administer it to a large, representative reference population and then scale all raw scores so that the median result equals precisely 100.
This design is shared by the major clinical assessment tools. The Wechsler Adult Intelligence Scale (WAIS), now in its fifth edition (WAIS-V, released by Pearson in 2024 with updated US norms), uses mean 100 and SD 15. So does the Stanford-Binet, 5th edition. The American Psychological Association provides broader context on what intelligence tests measure and their appropriate use.
Why 100 Is Always the Midpoint
The number 100 has no intrinsic meaning attached to raw reasoning ability. It is a reference anchor. If you administered the same test to a population with different educational backgrounds, the same raw score would yield a different scaled result. The 100-point anchor is only meaningful within the norm group the test was designed around. This is why percentile framing matters more than the headline number.
How Norming Calibrates the Scale
Standardisation means the test is scored relative to a norming sample, not on an absolute scale. Norming samples typically include 2,000 to 4,000 participants stratified by age, education level, gender, and geographic representation. All score reports you receive are relative to that sample — not to all humans who have ever existed, or even all adults in your country today.
The 15-Point Standard Deviation
Most major IQ tests use a standard deviation (SD) of 15. This sets the spread of the distribution around the 100 midpoint:
One SD above average = 115 (approximately 84th percentile)
One SD below average = 85 (approximately 16th percentile)
Two SDs above = 130 (approximately 98th percentile)
Two SDs below = 70 (approximately 2nd percentile)
Note: The Cattell Culture Fair test uses SD 24 rather than 15, so raw scores from that instrument cannot be directly compared to WAIS results without conversion.
What About the 68-95-99.7 Rule?
In any normal distribution, approximately 68% of values fall within one SD of the mean, 95% within two SDs, and 99.7% within three SDs. Applied to IQ: 68% of the reference population scores between 85 and 115; 95% between 70 and 130; and 99.7% between 55 and 145. These numbers appear in every IQ explainer because they follow directly from normal distribution mathematics, not from IQ research specifically.
IQ Score Ranges: Every Band With Percentile Context
Expand each band to see percentile data and practical interpretation notes. For the complete seven-band clinical classification system used by the WAIS-IV, see the complete IQ score ranges guide.
130+Very Superior+2 SD or above
98th percentile and above
Percentile: 98th percentile and abovePopulation share: Approx. 2% of the reference population
This range sits two or more standard deviations above the mean. The WAIS-IV labels it Very Superior. It is the threshold for most high-IQ society admission, though formal membership requires a supervised clinical assessment, not an online test result. Score variance is proportionally larger at extremes — a 5-point swing here shifts your percentile less than the same change near the centre.
120 – 129Superior+1.33 to +1.93 SD
91st to 97th percentile
Percentile: 91st to 97th percentilePopulation share: Approx. 7% of the reference population
The Superior band represents clearly above-average performance. A result here means you outperformed roughly 9 in 10 people in the comparison group. Improvement within this band is mostly about eliminating specific error types — rushing spatial rotation items, misreading inductive patterns on later questions — rather than wholesale strategy changes.
110 – 119High Average+0.67 to +1.27 SD
75th to 90th percentile
Percentile: 75th to 90th percentilePopulation share: Approx. 16% of the reference population
What most people informally call a strong score. Consistent above-median performance without reaching the rarer Superior band. Most people in this range benefit from targeted practice on the specific item types they find hardest — typically spatial rotation and inductive matrix tasks — rather than general preparation.
90 – 109Average-0.67 to +0.60 SD
25th to 73rd percentile
Percentile: 25th to 73rd percentilePopulation share: Approx. 50% of the reference population
The clinical average band as defined by WAIS-IV. The largest single slice of any normed distribution. A score here is solidly within the normal range by every major clinical or research standard. The gap between 90 and 109 is primarily about pacing, error reduction, and item-type familiarity — not some fixed cognitive ceiling.
80 – 89Low Average-1.33 to -0.73 SD
9th to 24th percentile
Percentile: 9th to 24th percentilePopulation share: Approx. 16% of the reference population
Still within the broad normal range, but below the clinical average band. Environmental factors — poor test conditions, unfamiliarity with format, fatigue — can produce results here even for people whose genuine reasoning baseline sits higher. A second clean-session attempt is warranted before drawing conclusions.
70 – 79Borderline-2 to -1.4 SD
2nd to 8th percentile
Percentile: 2nd to 8th percentilePopulation share: Approx. 7% of the reference population
The borderline range sits at the lower tail of the distribution. Clinical frameworks use this band as a decision threshold for further assessment, not as a diagnostic category on its own. A single online IQ-style result in this range should be verified under controlled conditions before any conclusions are drawn.
Percentile approximations based on a standard normal distribution with mean 100 and SD 15. Actual values vary by test edition and norming sample.
The Flynn Effect: Why the Average IQ Score Has Shifted Over Time
The average IQ score does not stay fixed across generations. Raw test scores rose substantially across many countries over the 20th century — a trend documented by researcher James Flynn and now known as the Flynn Effect. The average rate of gain was roughly 3 IQ points per decade across many Western countries between the 1930s and the 1990s — meaning a person who scored exactly average in 1930 would score noticeably below average on a 1990 normed version of the same items.
Estimated cumulative raw score gain (Western countries, relative to 1930 baseline)
1930
Baseline
1950
+6 pts
1970
+12 pts
1990
+18 pts (peak gain era)
2010
~flat / slight decline
2026
Modest reversal in several countries
Approximate estimates based on Flynn (1984, 1987), Teasdale & Owen (2008), and Pietschnig & Voracek (2015). Actual gains varied significantly by country and subtest type.
What the Research Data Shows
Studies from Scandinavia, the UK, the Netherlands, and the United States documented consistent gains across vocabulary, spatial reasoning, and abstract pattern tasks. The gains were not uniform across all subtests — fluid reasoning tasks (the kind measured by Raven-style matrix problems) showed some of the largest improvements, which is relevant context when interpreting any Raven-style test result today.
Causes Researchers Attribute to the Rise
Broader access to formal education and abstract-style problem solving
Reduced exposure to cognitive-impairing environmental toxins (notably lead)
Increased familiarity with standardised test formats and visual reasoning tasks
General enrichment of cognitively stimulating environments
The Reverse Flynn Effect: Declining Scores Since the 1990s
More recent data from several high-income countries — notably Norway, Denmark, and Finland — suggests the rising trend plateaued or reversed slightly from the late 1990s onward. A 2024 study in Intelligence (ScienceDirect) found inconsistent Flynn Effect patterns across cohorts from 2005 to 2024, with some subtests declining while others continued rising. A 2023 Northwestern University study using NAEP data found declining scores in verbal reasoning and mathematical comprehension among recent US cohorts. Researchers are still debating causes, with hypotheses ranging from shifts in educational approach to the cognitive effects of high-volume passive media consumption.
What This Means for Your Score Interpretation
Your score should be interpreted relative to the specific norm sample your test uses — not against historical averages or populations from different eras. A person who scored 100 on a 1970-normed version of the WAIS would likely score noticeably lower on a 2024-normed version of the same items. This is why the WAIS-V (2024) required fresh norming rather than simply adopting WAIS-IV standards.
Average IQ Score by Country: Data and Caveats
Compiled datasets — most prominently from researchers Lynn and Vanhanen, updated by Becker (2019) and cross-referenced against PISA international education scores — report national average IQ estimates ranging from roughly 70 to 108. These figures require substantial methodological caution before drawing any conclusions. For a dedicated country-by-country breakdown, see the IQ score by country guide.
Selected National Estimates (2019 – 2025 compiled data)
Country
Estimated Avg. IQ
Data Notes
Hong Kong
107 – 108
Multiple standardised assessments
Singapore
107 – 108
National assessments, high PISA correlation
South Korea
106 – 107
Military and school data
Japan
106 – 107
National standardised samples
Taiwan
104 – 106
National data, PISA-correlated
China (mainland)
103 – 106
Wide variation by region and sample
Finland
100 – 101
Military conscript data
Germany
99 – 100
National school and military samples
United Kingdom
99 – 100
Multiple standardised assessments
United States
97 – 98
WAIS normative data, NAEP cross-reference
Estimates derived from Lynn & Vanhanen, Becker (2019), and PISA cross-references. Confidence intervals are typically ±5 points or wider. These are orientation estimates, not clinically normed figures.
Key Methodological Limitations
Sample quality varies widely. Some national estimates derive from a single study with fewer than 100 participants. Wide confidence intervals make country rankings essentially unreliable.
Different tests across countries. Scores from different instruments with different norm bases are not directly comparable. A score of 100 on one national test is not equivalent to 100 on another.
Test familiarity gaps. Populations with less exposure to standardised test formats score lower on test-taking skill independently of actual reasoning ability.
Interpolated values. For some countries, researchers estimated figures from neighbouring country data rather than primary collection.
Environmental vs. Genetic Explanations
The pattern of higher average scores in high-income countries with well-resourced education systems is consistent with environmental explanations — not with fixed biological differences between populations. Research from Nisbett et al. (2012) in American Psychologist is among the most comprehensive summaries of what is actually understood about group-level score variation: environmental factors including education, nutrition, and test familiarity explain the preponderance of documented variation.
What Your Score Means Relative to the Average
If you have already taken an IQ-style assessment, the most useful question is not “is this above or below average?” — it is “what percentile does this correspond to within the norm group, and how reliably does this single session reflect my actual reasoning baseline?”
Percentile Position Matters More Than Raw Numbers
A raw score of 112 on one test might correspond to the 79th percentile. The same number on a different test normed on a different population might map to a higher or lower percentile. The raw score alone is not self-interpreting — it only carries meaning when attached to the percentile position it represents within the norm sample. For a visual reference of where different scores land, see the IQ score chart.
IQ Score
Approx. Percentile
Classification
1 in X
145
99.9th
Exceptionally Rare
1 in 1,000
130
98th
Very Superior
1 in 50
120
91st
Superior
1 in 11
115
84th
High Average
1 in 6
110
75th
High Average
1 in 4
100
50th
Average
1 in 2
90
25th
Average
1 in 4 (below)
85
16th
Low Average
1 in 6 (below)
70
2nd
Borderline
1 in 50 (below)
Based on standard normal distribution, mean 100, SD 15. Values are approximations.
Why a Single Session Is Only an Estimate
Clinical assessment frameworks such as the WAIS use test-retest reliability studies to quantify score variance between administrations. Even under well-controlled clinical conditions, score variance of 5 to 10 points between sessions is common. Online IQ-style assessments conducted without standardised conditions carry higher variance. A single result should be treated as directional evidence within a range, not as a precise fixed value. For high-stakes decisions, a supervised clinical assessment is the appropriate tool.
For an explanation of what constitutes a high result and how to interpret performance above 110, see the high IQ score guide.
Conditions That Affect Your Test Result
Sleep
Sleep deprivation of even one night can reduce performance by 5 to 10 points. Testing rested is not optional.
Noise and distraction
Background noise and interruptions measurably increase error rates on timed reasoning tasks.
Test familiarity
First-time exposure to matrix-style reasoning reduces performance by a real but quantifiable amount that decreases with practice.
Anxiety and time pressure
Test anxiety compresses the working memory capacity needed for multi-step spatial reasoning items.
Interactive Calculator
Where Does Your Score Fall?
Drag the slider or type a score to see your percentile and classification band.
557085100115130145
Percentile
50th
Higher than 50% of the reference population
Classification
Average
25th to 75th percentile. The core of the bell curve.
Distance from mean
0
Points from the 100 midpoint (0.00 SD)
Lowest trackedAverage (100)Highest tracked
Want to know your actual IQ? IQMog uses a fixed 20-item Raven-style test to produce your IQ estimate and percentile band. Start the free test to compare your real result.
Percentile calculated using a normal distribution with mean 100 and SD 15. Actual values vary by test and norming sample.
🧩
20 Raven-style questions
Fixed item set. Every person sees the same questions, making your percentile directly comparable.
⏱
About 20 minutes
No strict time limit. Accuracy matters more than speed. Your report is prepared after completion.
📊
IQ estimate + percentile
Your score, band, and cognitive profile are unlocked in a single report after you complete the test.
Frequently Asked Questions
Targeting the most common questions about average IQ scores, percentiles, and test interpretation.
What is the average IQ score worldwide?
By statistical design, the average IQ score on any properly normed test is 100. This is not an observed midpoint of human reasoning — it is a deliberate calibration decision. Researchers set the reference population median to 100 during norming. A 2025 international dataset of over 1.2 million participants who took an identical online test found a global mean consistent with this 100-point anchor, though national figures ranged from roughly 70 to 108 depending on education infrastructure and sampling method.
Is a score of 100 really average?
Yes, within the population a test was normed on. A score of 100 means you performed at the 50th percentile of that reference group — exactly half scored higher and half scored lower. The important caveat is that norms vary by test edition, country, and the year of norming. The WAIS-V released in 2024 uses updated US norms, meaning raw scores that would have produced 100 on the WAIS-IV produce a different scaled score today.
What IQ score is considered above average?
Most scoring frameworks place above-average performance at 110 or higher, which corresponds to roughly the 75th percentile. Scores of 120 or above reach approximately the 91st percentile and are classified as Superior by WAIS-IV criteria. A score of 130 or above sits at the 98th percentile — the threshold used by most high-IQ societies, including Mensa International — though online assessments are not accepted as formal evidence for admission.
Why does the average IQ score keep rising over time?
This trend is called the Flynn Effect, after researcher James Flynn who documented it in 1984. Raw IQ scores rose by roughly 3 points per decade across many Western countries from the 1930s through the 1990s. Researchers attribute the gains to improved education access, better childhood nutrition, reduced lead exposure, and increasing familiarity with abstract test formats. More recent data from Norway, Denmark, and Finland suggests the rise has plateaued or slightly reversed since the late 1990s, a pattern some researchers link to changes in educational approach and media habits.
Can my IQ score change from test to test?
Yes. A single IQ-style test result is a performance snapshot, not a fixed trait. Factors including sleep quality, testing environment, noise, test familiarity, and general stress can shift a score by 10 to 15 points or more between sessions. Even under well-controlled clinical conditions, WAIS test-retest studies show score variance of 5 to 10 points. Treat your first result as a directional baseline, not a permanent label. A second clean-session attempt — quiet room, rested, no interruptions — provides more reliable evidence than a single data point.
IQMog
Find Out Exactly Where You Stand Against the Average
Take the free 20-item Raven-style test, then unlock your IQ estimate, percentile band, and full report after completion. Every person sees the same fixed item set, so your result is directly comparable. No email required to start.