Frequently Asked Questions

Psychometrics is the scientific discipline concerned with the measurement of psychological attributes. It aims to quantify otherwise invisible traits such as intelligence, personality, motivation, or aptitude through structured assessments and statistical models.

Introduction & Overview

What is IQ?

IQ (Intelligence Quotient) is a score you get from a set of varying cognitive tests which compare your mental skills with other people your age. Test makers set the average score at 100, with one standard deviation being 15 points; most people fall between about 85 and 115.

IQ tests don’t measure just one thing. They tap many abilities-for instance, spotting patterns quickly (fluid intelligence), recalling learned words and facts (crystallized intelligence), and keeping several digits in mind at once (working memory). These are only a few of the many skills the tests sample. Because most mental skills overlap, people who score high in one area tend to score high in others. Psychologists call this overall overlap "g" for general intelligence. However, this overlap is not just a statistical phenomenon; it has deep roots and empirical standing in neuroscience, biology, genetics, and psychology.

IQ vs. Intelligence

IQ is a statistical proxy for the g factor, not a complete map of "intelligence". Also, IQ is not a perfect stand-in for "intelligence' if you define intelligence broadly or differently. However, IQ is considered the strongest single quantitative measure of cognitive ability-especially in predicting certain life outcomes (academic achievement, job performance, etc.).

g-Factor, g-Loading, and Reliability

• g factor (also known as general intelligence) is the common variance across a range of cognitive tasks. Statistically, it’s the first principal component extracted when psychologists run a factor analysis on many subtests, often accounting for 30-50% of the total score differences among individuals. Conceptually, g sits at the top of hierarchical models (e.g., Cattell-Horn-Carroll), predicting broad abilities such as verbal comprehension and fluid reasoning, and it correlates modestly yet consistently with real-world outcomes like educational attainment, job performance, and even health indicators. Further Reading

• g-loading is the degree to which a test or subtest correlates with the g factor or general intelligence. A higher g loading means the task is drawing heavily on general intelligence, and figures above 0.8 are generally considered to be great. g-loadings are often derived from a factor analysis and high g-loadings are prized when the goal is to estimate overall cognitive ability efficiently.

• Reliability is the consistency of test scores across time, forms, or item samples, quantified by coefficients such as test-retest, split-half, or Cronbach’s alpha. Full-scale IQ batteries typically aim for reliabilities above 0.90, yielding a small Standard Error of Measurement (SEM) so an observed score of 110, for example, likely reflects a “true” score within ±3-4 points. Without high reliability, even a strongly g-loaded test cannot be trusted for diagnoses, research comparisons, or tracking developmental change, because score swings could reflect measurement noise rather than real differences in ability.

Full Scale IQ (FSIQ) and Other Indices

FSIQ is formed by aggregating scores from a diversified set of cognitive tasks (reasoning, memory, verbal comprehension, visual-spatial analysis, processing speed, and so on) each chosen for its strong loading on g.

Various tasks combining to distill g

Because the tasks sample different mental operations, their task-specific noise tends to cancel out when combined, while the shared variance (the influence of g) accumulates. This makes FSIQ the most reliable single summary of g and overall cognitive ability: it minimizes measurement error and maximizes predictive validity for broad outcomes such as academic learning, occupational training, and life-course problem-solving.

The most common subtests included in FSIQ tests fall under the following broad factors:

• FRI - Fluid Reasoning Index: gauges how well you solve novel problems and detect patterns without relying on prior knowledge.
• VCI - Verbal Comprehension Index: captures your grasp of vocabulary, verbal reasoning, and general knowledge expressed through language.
• VSI - Visual Spatial Index: measures the ability to perceive, analyze, and mentally manipulate visual-spatial relationships.
• QRI - Quantitative Reasoning Index: assesses understanding of numerical concepts and effectiveness at mathematical problem-solving.
• WMI - Working Memory Index: reflects how efficiently you can hold and transform information in immediate awareness.
• PSI - Processing Speed Index: times the speed and accuracy with which you carry out simple, routine cognitive tasks involving visual information.

Genetic vs. Environmental Influences on IQ

IQ is largely heritable (estimates range from ~50 % in childhood to ~80 % or higher in adulthood). The Wilson Effect describes how genetic influence on intelligence grows stronger with age. Environmental effects (nutrition, early childhood education, extreme stress, etc.) can influence the phenotypic expression of IQ, but large permanent changes to “true g” in adulthood are unlikely.

Can I Improve My IQ?

No reliable method is known to permanently increase g itself for healthy adults. Practice on certain item types can raise your test performance on those items, but that generally does not reflect a genuine rise in general intelligence. Key factors that support you reaching your genetic potential include:

• Adequate nutrition and sleep
• Exercise and overall brain health
• Avoiding or managing depression, anxiety, and ADHD-related inattentiveness

Age Effects and the Wilson Effect

Childhood IQ can be quite variable, with environment having a larger relative influence.

Heritability with Age (Wilson Effect)

By late adolescence or early adulthood, heritability is at its highest, and your measured IQ tends to stabilize.

The effect of aging on various cognitive factors

In older adulthood, some indices (e.g., processing speed) may decline, while crystallized abilities (verbal knowledge) often remain stable or even improve until mid or late adulthood.

Decomposition of the variance in FSIQ over time

As people approach later maturity, the impact of genetics takes over, reaching an asymptote of ~0.80 at 18-20 years old and remaining stable going forward. As age progresses, genetic influence on intelligence strengthens while environmental impact diminishes and your childhood scores may have been impacted by this.

This may also explain the "gifted kid burnout" syndrome. Just as some were the tallest in their class as kids but stopped growing and are average height in adulthood, those who were "gifted" as kids may struggle to meet those same expectations as adults. However, the inverse may also be true, analogous to growth spurts.

---

Which Tests Are Best?

Are Online Tests Accurate?

Most online tests that can be found on Google are not good, unless they have statistical validation. However, there are trustworthy tests linked in this subreddit's resources tab, ranked by g-loading and other factors. Online administrations of professional tests, such as those proctored through Discord, are also accurate as long as the proctor follows the manual exactly and the examinee is in an optimal state, comfortable environment, and a strong connection. However, diagnoses from scores should be taken with a far larger grain of salt if they're by non-professionals. If you took a test in real life, ask your psychologist to interpret before posting to the subreddit, because they can do a better job than with our limited knowledge of your situation.

Recommended Online Tests & Rankings

Below is the general consensus among experienced members of this community. For a more comprehensive list, please check out the following resources list.

	FSIQ Tests	VCI Tests	FRI Tests	VSI Tests	QRI Tests	WMI Tests	PSI Tests
S Tier	Old GRE, Old SAT	Old GRE V, Old SAT V	Old GRE A	PAT	Old GRE Q, Old SAT M
A Tier	AGCT, AGCT-E	MAT, VAT-R, CAIT VCI	1926 SAT FRI, CAIT FW	SAE	SMART	Digit Span, Spatial Span, Corsi	WAIS-III Coding
B Tier	CAIT, 1926 SAT	CMT-A, CMT-T, IAW, 1926 SAT KN+VR	JCTI, WN, JCFS, D-48, Tutui R	CAIT VSI, MRT	1926 SAT QRI	Running Digit Span	CAIT SS
C Tier (and below)	Anything Else	Anything Else	Anything Else	Anything Else	Anything Else	Anything Else	Anything Else

Details on Recommended Tests

• Old SAT (pre-1995 recentering) and Old GRE (pre-2011) are frequently recommended in this subreddit as some of the best free measures of FSIQ. They have:
  • Extremely high g-loadings (around 0.92-0.93).
  • Large normative samples, giving them excellent predictive validity and stable high-end ceilings.
  • Well-documented scoring tables and correlation data with official IQ measures.
• AGCT (Army General Classification Test) and its extended (amateur) version (AGCT-E) are also solid, old standardized measures used historically by the US military with sample sizes in the millions as well.
• CAIT (Comprehensive Adult Intelligence Test) is popular on this subreddit for measuring a broader Full Scale IQ (FSIQ) across multiple indices.
• 1926 SAT is an interesting historical measure with a very high ceiling for fluid reasoning, though it is older and can feel a bit outdated.

Why Modern SAT/GRE Are Weaker for IQ

The modern SAT/GRE are not good IQ tests, being more susceptible to practice effect and largely focusing on knowledge gained from a solid education, as opposed to innate intelligence. The reasons that the Old SAT is not administered are numerous and complicated. Here are a few reasons that I believe to be the driving factors (though please note that these may be oversimplified and not 100 % factual):

• An increase in the number of people attending college. The College Board needed a test that catered more towards the average (and below-average) students, so they decided to re-center the test, making the average person get a higher score and reducing the ceiling of the test percentile-wise.
• Changing technology. The handheld calculator (and in more recent times, the internet) has completely transformed high school education, and it was important that the College Board took this into account.
• Anti-Asian and anti-Jewish racism. These two racial groups generally outperform all others on the SAT, and there may have been a lot of concern amongst white elites about their spots being taken by poor non-white students. Thus, they created a test with a lower ceiling that could be practiced for more easily.
• Progressivism and changing political ideals. Most of the highest scorers on the SAT were male (due to a theory called the Greater Male Variability Hypothesis), and most of the lowest scorers were Black, Native American, and Latino. Because these score differences reflect the exact same score differences in IQ-something seen as an innate trait-it is possible that this was done to avoid anti-discrimination lawsuits (though these differences have still emerged on the modern SAT, albeit to an arguably somewhat smaller extent). There was also an increasing emphasis on top universities accepting an equal number of men and women, utilizing affirmative action to create a more racially representative class, and adopting an overall more “holistic” admissions approach, as opposed to one that emphasized a singular innate trait so heavily.
• The publication of The Bell Curve in 1994. This brought attention to the SAT being an IQ test with the aforementioned discrepancies between demographic groups.

There are a lot more factors at play, and whether or not these changes have been good is up for debate, but this is what we believe to be the gist of the issue.

---

Test Design & Common Questions

How come these tests have vocabulary and “trivia” questions? Doesn’t that defeat the point of IQ? Why isn’t matrix reasoning on the SAT?

You are missing the entire point of an IQ test. The point of an IQ test is to serve as a proxy for ‘g,’ or general intelligence. G is a latent trait, meaning that it cannot be directly observed or measured with 100 % certainty. So, instead, people have devised tests that approximate it incredibly well using a process called factor analysis. Verbal tests generally have the highest g-loading because they consist of words/facts that almost everyone has been exposed to at some point in their lifetime-i.e., they are not arbitrarily selected. Professionally developed verbal tests often take months, if not years, to be completed and are not just an accumulation of the author’s favorite facts or cool-sounding words. They are specifically meant to test words that everyone will have been exposed to (even with differing levels of access to quality education) and facts belonging to the Western canon that would be covered in a typical American K-12 education. Additionally, verbal tests often test more than mere recall and/or the ability to list off the dictionary definition of a word. They also focus on your ability to use/understand the word contextually, relate the word to other concepts, reason with the word, etc. If you think about it, it makes a lot of sense that people who remember and reason more accurately with words/facts that everyone has been exposed to are, on average, going to be more intelligent than those who remember and reason with them more inaccurately.

How come there is math on this test? Will it be valid for me as a 77-year-old Rwandan who has never seen Hindu-Arabinumerals?

There is math on the test because it is something that the average American high schooler or college student will have been exposed to. Additionally, our everyday lives are filled with quantitative reasoning, and humans appear to generally be innately quantitative individuals. Your ability to engage in mathematical reasoning is imperative to success in the modern world and thus captures an important part of intelligence for the vast majority of people. The math questions contained within tests on the sub (SAT, GRE, AGCT, SMART, etc.) are all able to be solved with extremely basic arithmetic, algebra, and geometry, making them incredibly predictive measures of fluid reasoning for the populations that they were intended to measure (as opposed to crystallized measures of math education and/or achievement). That being said, if you are significantly older/younger than the average test-taker, received a math education outside of the US, have little to no math education, did math olympiads, have dyscalculia, etc., these sorts of factors may reduce the accuracy of QRI tests at measuring your g, leading to marginally inflated or deflated results. However, they will produce an accurate measure of how well you use quantitative reasoning compared to the average person on a daily basis, which is arguably more useful to you as a test-taker than knowing your exact 1.00 g-loaded IQ score.

I’m non-native; can I translate the words? Is my IQ going to be accurate without VCI?

No, you cannot translate the words. They may translate to words that are either more or less common in your native language, not to mention that concepts/words that are common in one culture might be far less common in another, and vice versa. This dilutes the ability of the verbal test to accurately measure your intelligence. Considering how important verbal intelligence is, your IQ is not going to be accurate without VCI. The good news is that you can get quite close through non-verbal tests (as well as some tests where only very basic English proficiency is necessary).

---

Fairness, Bias & Neurodivergence

Are IQ Tests Racist or Sexist?

A common misconception seems to be that because IQ tests don’t produce equal measurements for everyone, they are inherently biased and flawed instruments. However, this is not necessarily the case, as there could be biological differences between different groups, and/or the IQ tests could be measuring various societal forces that are at play. Regardless, they seem to be an accurate measure of daily intellectual functioning or the way that g manifests in the real world, which, once again, is arguably more important to be measuring than g itself. That being said, all mathematical evidence points to IQ tests being extremely accurate measures of g, so while population-based differences in g may be rooted in biological, social, cultural, socioeconomic, or political differences, it is likely that innate, unchangeable differences in g still exist and are being accurately picked up on by IQ tests, even if they will not exist forever as society and the gene pool change. So no, IQ tests themselves are not racist, though they may pick up on racism and sexism (or they may just be picking up on biological differences, who knows), but this just makes them even more accurate/meaningful and increases their predictive validity.

How can IQ tests be accurate for populations like neurodivergent people?

IQ tests may not be as accurate at capturing a neurodivergent person’s true genetic potential if they are untreated and unmedicated. However, they will still be equally as accurate at assessing that person’s intellectual functioning on a daily basis when compared with others. If a person cannot reach their genetic potential when trying their best on an IQ test, it seems highly unlikely (though perhaps not impossible) that they would magically reach their full genetic potential in real-world endeavors. Something else to note is that studies have shown IQ testing for subjects with autism and ADHD to be measurement invariant, meaning that these tests are measuring the same construct of intelligence with equal accuracy for neurodivergent testees when compared with neurotypical testees. Additionally, IQ tests can be useful for identifying neurodivergence or the ways that g manifests in a neurodivergent person’s life.

If IQ may have issues measuring these various populations, why aren’t more specialized norming samples done?

It does not make practical sense to. If you are 77, an IQ test isn’t going to hold as much value for you compared to an 18-year-old. If you’re a woman in Somalia, an American IQ test may not be super predictive or accurate in your society. If you have an IQ of 160, people will only meet two other people as smart as you in their entire lifetime. This is not to say that you don’t matter or that we shouldn’t strive to produce more accurate measures of intelligence for everyone, but it simply isn’t worth it to invest the time or money into developing accurate tests for these populations from the perspective of Western testing companies.

---

Interpreting Scores & Variability

IQ Score Distributions and Rarity

IQ Range	Typical Classification	Rarity (SD = 15, M = 100)	How Many Exist (assumes 8.1 billion people)
140 +	Highly Advanced	1 in 261 people upward	30.8 million people
130-140	Very Superior/Gifted/Very Advanced	1 in 44 to 1 in 261 people	159 million people
120-129	Superior/Very High	1 in 11 people to 1 in 38 people	560 million people
110-119	High Average	1 in 4 people to 1 in 10 people	1.292 billion people
90-109	Average	1 in 4 people to 1 in 4 people	4.03 billion people
80-89	Low Average	1 in 11 people to 1 in 4	1.292 billion people
70-79	Borderline impaired or delayed	1 in 44 people to 1 in 12	560 million people
55-69	Mildly Impaired or Delayed/Very Low	1 in 741 people to 1 in 52	174 million people
40-54	Moderately Impaired or delayed/Extremely Low	1 in 31 560 to 1 in 924	10.3 million people

Calculated assuming a perfect distribution

Why Do My Scores Vary Between Tests? (Regression to the Mean)

No, the tests are still accurate, but you are just succumbing to a common effect known as regression to the mean. In any population of high IQ people, there are bound to be some false positives, so when these people are retested, they usually score lower than the original mean. One example is Mensa, where the members were found to actually average an IQ of closer to 120 SD 15, despite qualification to the organization being a score of 130 IQ SD 15 on an IQ test. The better the test, the lower the drop-off, and the further from the mean, the more regression there is. Let’s use the Old SAT as an example. According to the “Big ‘g’ Estimator” on CognitiveMetrics, someone who scores 130 on the Old SAT most likely has an actual IQ of 128. Likewise, someone who scores 160 most likely has an IQ of 156. Now, let’s look at the CAIT. Someone who scores 130 on it most likely has an actual IQ of 126. Finally, let’s look at Raven’s 2. Someone who scores 130 on it most likely has an actual IQ of 121. However, things can go the opposite way, too, because it is abnormal to never succumb to regression to the mean. Thus, someone who scores 130 on the Old SAT, CAIT, and Raven’s 2 most likely has an actual IQ of 131. Another thing to note is that the g-loading often drops off the further from the test’s mean you get. This means that more non-g factors, such as sleep or motivation, may be at play, increasing the variation in your scores. It is perfectly normal to score 140 on one form and 130 on another. If you take more forms, you will probably find that your average lies somewhere in between the two.

I scored 78 but am highly successful in real life, despite my score. Why is that?

IQ tests can be predictive of a number of life outcomes associated with “success.” However, even though they are often the largest predictive factor, they are never the only predictive factor, and oftentimes they may not even account for the majority of the variance. It is possible that you are really gifted in non-g-related factors that also predict success well (or ones that don’t have a clear correlation but can lead to success if used appropriately), or that you are simply a large statistical outlier, which is just a thing that happens.

Why does someone I know who scored 145 on a test seem smarter than someone who scored 160 on a test?

IQ becomes less accurate the further from the mean you get. Additionally, intelligence becomes varied as the chances of someone scoring 150 in one index are higher than a person scoring 150 in six indices. There are probably also biological factors that lead more intelligent people to have more relative strengths and weaknesses than your average 100 IQ person. So, there are a couple of possibilities. One possibility is that they took different tests. A 145 on JCTI and a 160 on MAT are measuring different things, and this could be playing to a person’s strengths or weaknesses, even if both testees have similar FSIQs. This is even true of FSIQ tests; a 160 on CAIT and a 160 on GRE are still measuring slightly different things and may play to individual strengths or weaknesses. Another possibility is, of course, that the test just produced slightly inflated results for one and slightly deflated results for the other due to some small non-g-related factors and/or the ceiling effect. Finally, there is the lack of accurate norming. Most professional tests, such as the SB-V or the WAIS-IV, do not have 30 000 people in their norming sample. So, scores of 160 are extrapolated based on the norming samples that they do have. However, this may lead the rarity of a score of 160 to be over- or understated. While it is true that, on average, someone who scores 160 on the WAIS-IV is more intelligent and will have better life outcomes than someone who scores 145, it is hard to actually tell by how much and how meaningful this prediction is. At this point, IQ scores often become largely relative. It is clear that a score of 160 is better than a score of 145, but is it actually 10 times rarer, especially when accounting for things like log-normality, ceiling effect, regression to the mean, and an inadequate sample size? This is why, if you are scoring >130, your best bet for an accurate score, reflective of real-life percentiles, is going to be the Old SAT/GRE, as these tests have far larger norming samples and higher levels of predictive validity for gifted people than SB-V/WAIS.

“190 IQ” Claims and Ratio IQs

Back in the day, IQ tests used to be reported in ratio scores, leading to many inflated results, especially for children. Additionally, some countries, such as the UK, may report scores in a different SD, such as SD 24. Finally, a lot of people just lie, especially on the internet or in spaces meant for the “gifted,” which more often than not just attract lots of mentally ill LARPers. If someone reports an IQ score greater than 160 SD 15, there is a 99.9 + % chance that they are full of [it].

---

Indices and Subtests Explained

FRI, VCI, VSI, QRI, WMI, PSI

Each of these indices are measures of different subsets of g, as defined by the Cattell-Horn-Carroll Model of Intelligence. These can be composited to calculate GAI or FSIQ. FRI stands for Fluid Reasoning, VCI stands for Verbal Comprehension, VSI stands for Visual Spatial, and QRI stands for Quantitative Reasoning Index. CPI stands for Cognitive Proficiency and usually consists of WMI (Working Memory Index) and PSI (Processing Speed Index).

• FRI (Fluid Reasoning Index): Ability to detect and apply logical patterns, often with novel, nonverbal content.
• VCI (Verbal Comprehension Index): Vocabulary, verbal reasoning, concept formation.
• VSI (Visual Spatial Index): Spatial manipulation, such as block design or puzzles.
• QRI (Quantitative Reasoning Index): Reasoning with numbers, relationships, and quantitative data.
• WMI (Working Memory Index): Holding and mentally manipulating information (digit span, spatial span).
• PSI (Processing Speed Index): Speed and accuracy in simple tasks (symbol coding, search).

Why Might WMI/CPI Scores Be Low?

It may be indicative of a neurobehavioral disorder, such as ADHD or ADD. There is no way to tell for sure, so professional input is greatly valued. Most studies indicate that ADHD itself does not lower IQ, but it will impact your cognitive function. Additionally, there is currently no direct evidence suggesting OCD impacts IQ, but the interference itself with cognitive function will affect performance on tasks requiring attention. Depression, on the other hand, has been proven to affect cognitive function oftentimes in CPI-related areas such as memory. It has also been noted to have an overall permeating effect on executive function, but to quantify exactly how much it affects a score or your exact IQ is a pointless endeavor.

---

Combining & Averaging Scores

The “Big g” Compositator

If you’ve taken multiple highly g-loaded tests (e.g., Old SAT, CAIT, Raven’s, etc.), you can use the “Big ‘g’ Estimator” on CognitiveMetrics to factor in each test’s correlation with g. It will generate a composite estimate of your FSIQ.

Dealing with Differing Standard Deviations

Remember that some tests (e.g., older Stanford-Binet) use SD = 16, or Cattell scales might use SD = 24. Always convert to a “common” scale (usually mean 100, SD 15) before you average or compare them.

---

Miscellaneous Questions

Genome-Wide Association Studies (GWAS)

Genome-Wide Association Studies (aka GWAS) analysis can sometimes be quite accurate. They work by identifying DNA variants that contribute to the variance in intelligence. Due to the strong genetic nature in IQ, GWAS can identify many single-nucleotide polymorphisms (SNPs) which have an effect on intelligence. The studies so far use a polygenic score which can be used to see how predictive it is of IQ. However, some studies find that polygenic scores of intelligence can predict only around 4 to 10.6 % of the variance in intelligence, so it isn't perfect. It is also cautioned to be careful with where you upload or share your data.

MBTI, Big Five, and Other Personality Measures

This is an interesting question considering MBTI has been tested or inquired about in many scientific articles. The consensus is that MBTI itself lacks scientific rigor and exists purely as a commercial vessel, but that does not mean the score you get cannot be reliable nor accurate if MBTI were to adhere to strict definitions and scoring is dependent on a sample size and calibration of results. Another alternative would be the Big-5 which is perceived more positively than MBTI in academia.

---

Tables & Data

Calculated g-Loadings of Major IQ Tests

Test	g-loading	Date Published
SB-V	0.96	2003
WAIS	0.94	1955
WAIS-III	0.93	1997
SB-IV	0.93	1986
SAT	0.93	1974-1994
GRE	0.92	1981-2001
WAIS-IV	0.92	2008
WJ-IV COG	0.91	2014
AGCT	0.92	1941
WISC-V	0.90	2014
WISC-IV	0.90	2003
WAIS-R	0.90	1981
WISC-III	0.90	1991
WB	0.90	1951
WASI-II	0.86	2011
RIAS	0.86	2003

IQExams Tests g-Loadings

Test	g-loading	ωₕ	α	n at time of analysis	Type
HumanIQ^	0.747	0.680	0.863	2543	Spatial
High Range RT^	0.738	0.664	0.842	458	Spatial
Tero41^	0.733	0.655	0.869	1172	Spatial
LDSE^	0.724	0.638	0.925	166	Spatial
Logica Stella	0.719	0.630	0.858	1492	Spatial
Astrolab36	0.715	0.624	0.861	372	Spatial
Octagon	0.711	0.616	0.877	393	Spatial
Processor40^	0.689	0.578	0.882	500	Spatial
Matrix3×3	0.681	0.565	0.907	333	Spatial
Fuse	0.678	0.559	0.823	372	Spatial
Level	0.674	0.554	0.779	362	Spatial
Backspace	0.657	0.526	0.842	239	Spatial
HoudinIQ^	0.651	0.517	0.797	432	Spatial
ArithmetIQ	0.630	0.498	0.799	471	Numerical
EvolutionaryTS^	0.603	0.443	0.752	245	Spatial
Momentum	0.603	0.442	0.733	876	Spatial
PINOT40^	0.599	0.450	0.754	271	Numerical
PMA32E^	0.581	0.411	0.666	1333	Spatial
Spat Analogies^	0.578	0.407	0.735	512	Verbal
Analogix	0.390	0.209	0.650	289	Verbal
Average	0.675	0.563	0.813	12831

^ Distribution may not be normal; interpret with caution.
Results for all tests and tiers were calculated around 2022-01-08 to 2022-04-18 and are subject to change if more data are provided (likely not).

This Community's Average IQ Scores Across Various Tests

Test	g-loading	Mean	SD	n
AGCT	0.92	120	13	10,318
CAIT	0.85	123	16	7,838
SAT	0.93	126	12.5	4,017
SMART	0.84	133	13	472

Why are these scores significantly above average?

The selection effect is when certain characteristics or traits are overrepresented in the sample due to the selection process. In this case, people who are interested in taking IQ tests online tend to score high on them to begin with, since their high scores drive the initial interest. A similar selection effect can be seen with SMART. Since it is a difficult, quantitatively-focused test, people who score high on quant tests are more likely to follow through and take the whole test.