Can you measure high IQ reliably?

On this page we address the ever-recurring question: “Can extremely high IQ be measured reliably?” The simple answer is: the higher the IQ score, the less reliable it becomes with conventional statistical measurement methods, unless further adjustments are made. Several factors influence how accurate or meaningful the measurement is. Some perspectives:

Answer as someone with IQ 101–120: No, it can’t, and IQ says absolutely nothing about anything. The only thing that matters is EQ.

Answer as someone with IQ 121–130: No, it only works up to a certain level, maybe around 130 or at most 160. After that, we’re all the same.

Answer as a Mensan with IQ 131–144: No. But Mensans are smarter than the norm, and no one can be smarter than a Mensan. Therefore, there is no IQ higher than what is measured for Mensa. Anyone who says otherwise is just talking nonsense.

Answer as a Triple Nine Society member with IQ 145–160: At the higher percentiles, both reliability and validity become increasingly complex challenges. The dynamics of measurement shift, and traditional quantitative frameworks serve primarily as foundational tools. Classical Test Theory (CTT) is employed to estimate measurement error and reliability; Cronbach’s alpha and split-half techniques are applied to assess internal consistency; test–retest correlations are used to establish temporal stability; standard error of measurement (SEM) in conjunction with confidence intervals provides estimates of error margins; and z-transformations as well as Item Response Theory (IRT) contribute to norming procedures and to the identification of items that retain discriminatory power at elevated ability levels. At the extreme upper tail of the distribution, ceiling effects diminish informational value. Here, adjustments such as empirically weighting observations, addressing sample biases, conducting more rigorous item analyses, and even incorporating qualitative expert assessments may be warranted. In such contexts, strict adherence to the assumption of a normal distribution may be less appropriate.

Answer as a Mega Society member with IQ 171–171+: Mensa is like the neighborhood pub; Mega Society is like a three-star Michelin restaurant. Even though it is statistically difficult to determine which restaurants truly qualify as three-star Michelin, they do exist—for those who would rather enjoy a delicate amuse-bouche followed by a well-composed tasting menu than settle for a bowl of lumpy stew. Here there is a shift from what can be quantitatively verified to what can be qualitatively discerned.

Of course, even what is qualitative can be disputed. But the small group of people who have dined many times at three-star Michelin restaurants know better when they are served something that is not of that caliber, compared to someone who has only been to the local pub.

So, although the quality of Michelin restaurants is not measured statistically, they still occupy the outer edge of what a restaurant can deliver. The same applies to people: at the extremes there is a high concentration of individuals with traces of brilliance that are hard to capture statistically. For example, Mega Society today has around 40 members—a very small and selective group. Among those who have qualified, or who attempted and still performed at an exceptional level, are:

• Ed O. Thorp
  – Mathematically proved that blackjack could be beaten through card counting and popularized how to “beat the house” (Beat the Dealer), university lecturer in mathematics, co-inventor of the first wearable computer, pioneer in arbitrage (Beat the Market), billionaire, etc.

• Marilyn vos Savant
  – Guinness World Record holder for the highest recorded IQ, author and columnist, popularized the Monty Hall problem in 1990, etc.

• Solomon W. Golomb
  – Mathematician, Franklin Medalist, pioneer in space communications, creator of the precursor to Tetris, etc.

• John H. Sununu
  – Former White House Chief of Staff, ex-Governor of New Hampshire, Ph.D. from MIT, etc.

• Cavan Cohoes
  – Mathematician, pilot, former professional athlete, instrumentalist, etc. (a modern Renaissance man).

• Georg Wikman
  – Mathematician, physicist, scientist, and entrepreneur, recipient of a “Lifetime Achievement Award,” first to commercialize Kan Jang, etc.

• Fredrik Ullén (narrowly missed the cut by one point)
  – Professor of neuroscience at Karolinska Institutet, concert pianist, director at the Max Planck Institute, member of Prometheus Society, etc.

Within this sphere, the numeric score is not the ultimate concern. It is recognized that measurement at such thresholds is inherently unreliable. Instead, the emphasis lies on the capacity to solve problems of exceptional difficulty—challenges few others can approach—and on achieving stringent qualifying thresholds despite obvious limitations in measurement.

It is also notable that considerable discourse surrounding IQ reflects less a genuine engagement with the construct than a defense of personal or group identity. Assertions that higher levels cannot be measured often function to preserve ego rather than to advance inquiry. This behavior is readily observed in professional settings, online forums, and informal conversations, including within societies such as Mensa. While not representative of all members, it is indicative of the individuals who express themselves in this manner.

As John Stuart Mill observed:

“It is better to be a human being dissatisfied than a pig satisfied; better to be Socrates dissatisfied than a fool satisfied. And if the fool, or the pig, is of a different opinion, it is only because they only know their own side of the question.”

In conclusion, the editorial team of this IQ blog notes substantial differences between societies with varying admission thresholds—such as Mensa, the Triple Nine Society, and Mega Society. Predictably, eccentric individuals may be found at all levels.

Want to know who belongs to the IQ elite? Read more:

The world’s highest IQs