I ran an IQ Contest at the Best University in Latin America (STEM Students)

Results and more information

Introduction

The Concurso de Acertijos en Exactas (Puzzles Contest in Exactas) was a project that I initiated and conducted completely independently. The prize for winning it was 10,000 Argentinean pesos; money that came out of my pocket.

The test consisted of 26 multiple-choice questions which showed visual patterns and by deciphering them you found the correct answer.

At the time I tried to hide the origin of the questions to prevent cheating, but it is no longer necessary. The truth is that I extracted them from the WAIS-IV, the most frequently used IQ test in the world. Its statistical reliability (reproducibility) and scientific validity (accuracy) are excellent.

The WAIS-IV contains multiple different sections. Some measure memory, others verbal comprehension or processing speed. I chose to use a subtest called "Matrix Reasoning" that measures abstract reasoning: a mental ability that I consider crucial for studying the exact sciences.

I came to discover this field of Psychology trying to know my cognitive capacity. Once my personal curiosity was satisfied, I wanted to test other people.

In a short time, I had no known people left to evaluate, so I looked for others in the University where I study Computer Science (which happens to be number one on the QS Ranking of the Latin America region). This time I was ambitious; I spent 100 hours looking for students that I randomly found around the 0+Infinity building. Anyone who went through the first year could participate.

Statistical Parameters

Given the sample size (252) and the population (7120), the confidence level would be 95% with a 6% margin of error; which means, in this case, that there is a 95% probability that the actual IQ is within ±6% of that detected by my experiment (mathematical proof).

Results

If you add up all those numbers you get 247, but wasn't it 252? Yes, it happens that I excluded other careers such as Geology and Oceanography (sorry). The participation was not enough to merit adding the data to the graph.

That students from Physics, Computer Science - Data Science, and Biology, respectively, made up the "68, 68, 69" circle was not planned, but I'm glad to have a fair representation of those majors. There is a rather smaller sample of Mathematics and Chemistry, though I think representative since fewer people study them.

In descending order their scores are set out:

Mathematics students have the highest average IQ, with Physics students in second place. In the middle, coinciding with the average of the College is Computer/Data Science. Chemistry and Biology occupy the last positions.

What do these numbers mean?

The ranking reveals that the average IQ of the College, and the median of all majors, are considered to be superior in a normal distribution.

The finding was both surprising and predictable. I expected to find many more intelligent people at the university than anywhere else (it is precisely one of the reasons why I study there), but not so many.

It seems obvious that just as you find especially athletic people going into Sports, disciplines which demand a high level of physical ability, you find especially intelligent people going into the Sciences, disciplines that demand a high level of mental ability.

To compare with the normal population of Argentina I made the following graphs:

An average person in Argentina scores 100; while an average student of the College of Exact Sciences of the UBA, 120, placing him in the 91% percentile (Top 9%) and making him 5 times rarer in intelligence.

Rare in a normal population, but "mediocre" in Exactas; where you find one (1) extremely intelligent person every 11 you come across: 24 times more frequent than expected in almost any other part of Argentina.

In this College, those with average intelligence are the "weirdos"; those with one which is superior under ordinary standards (>=98%) lose the privilege and go down in the IQ hierarchy.

Beyond the quiz results, different research has discovered and replicated an IQ difference by the level of education attained:

Note that Bachelor's Degree includes all majors. Those who graduate in Exact Sciences tend to score higher (by a difference of up to 10-15 points).

What we know about Human Intelligence

Intelligence is the ability of an agent to perceive or infer information, retain it as knowledge, and apply it in adaptive behaviors that allow him to function in his environment with greater success.

Fulfilling each of these functions requires different cognitive abilities, some of which have already been mentioned, such as memory or abstract reasoning.

Although there is considerable uncertainty on the subject, we can make a list, provisional until falsified, of what the current scientific evidence teaches us:

• Approximately 80% of intelligence is inherited from parents. The rest is determined by environmental factors (nutrition, prenatal conditions, etc.). The correlation is lower in childhood but higher in adulthood.

• A large proportion of genes have effects on the development of intelligence. None significantly on its own.

• Anatomical features that may determine intelligence are brain size and shape of frontal lobes, the amount of blood and chemical activity in frontal lobes, the total amount of gray matter in the brain, the overall thickness of the cortex, and the metabolic rate of glucose.

• Children with high scores on IQ tests tend to learn more of what is taught in school than their peers with lower scores. The correlation with their grades is about "0.50".

• The correlation of IQ with academic aptitude tests (e.g. SAT) is "0.82".

• The "multiple intelligence theory" lacks empirical evidence, and instead one is accepted which explains the existence of a general intelligence factor (g) that amalgamates a series of cognitive categories and abilities: it accumulates, historically, an enormous amount of strong confirmations in a great variety of social contexts.

• Increasing the intelligence of an adult is, so far, impossible.

• There are differences in intelligence between races, ethnic groups, and sexes; explained, in part, by socioeconomic differences. For example, Jewish people tend to score above average.

  This phenomenon was also replicated in my experiment, where the top four positions were occupied by individuals of Jewish descent: a grossly disproportionate representation given that they make up only 3% of the population of the City of Buenos Aires.

  What is even more surprising, 27% of Nobel Prize winners in Physics and 31% of the Nobel Prize in Medicine are Jewish, when they make up only 0.2% of the world's population: a disparity of up to 155 times the norm*.

  *This section is inspired by a genuine curiosity about a fact that impressed me. It is not my intention to justify hate speeches (which unfortunately exist and have claimed many innocent lives) or absurd anti-Semitic conspiracies.

To all the previous statements please add at the beginning an "It seems that...". Although they reflect the current scientific evidence on the subject, they are most likely not completely true.

The quintessential measure of intelligence is the IQ test: one of the most sophisticated psychological tests, subjected to the most exhaustive quality controls before publication.

Clinical psychologists generally consider IQ tests to have high statistical reliability. Reliability represents the consistency of measurement of a test, which, if high, produces similar scores when repeated.

Reliability and validity are distinct concepts. While reliability reflects reproducibility, validity refers to whether the test measures what it is intended to measure.

Despite some objections, clinical psychologists generally consider IQ scores to have sufficient scientific validity. These are not intended to give an absolute measure of intelligence, like the scale or grams concerning weight, but to rate the intellectual ability of an individual relative to others born in the same year: no matter the age group, the mean score is always set at 100.

Mainstream Science on Intelligence (1994), an editorial signed by 52 experts in the field, concludes:

Differences in intelligence exist, can be fairly measured,

are largely genetic, and influence life outcomes.

Epilogue

This text is the culmination of a long personal obsession with the subject of Human Intelligence (although Artificial too, that's why I study Computer Science) and the ways to measure it.

Please note the more than 20 linked scientific articles that support my arguments. Isaac Newton said: "If I have seen further it is by standing on the shoulders of giants".

You can read them for free as I resorted to Sci-Hub.se and LibGen.rs: great websites from which you can unlock access to almost any scientific article or book.

In this sentence, you will find the spreadsheet with the complete data report, the quiz puzzles, the test, and the answers. Click on the relevant words to find the specific resource.

More interesting data: