I said I would devote my first post of November to discussing my chronometric scores. Chronometrics is a method of measuring intelligence using reaction time (measured in milliseconds) to elementary cognitive tasks. The technique was pioneered by Sir Francis Galton in the 19th century but abandoned when Galton failed to find much if any correlation between measures of simple reaction time and school grades.

The idea was resurrected by psychologist Arthur Jensen in the late 20th century. With the advantage of modern statistical techniques and computer equipment, he discovered that there is indeed a moderate negative correlation between simple reaction time and IQ (the correlation is negative because the lower the reaction time, the faster the speed) especially when many attempts are average together. There’s an even higher negative correlation between choice reaction time and IQ. And there’s an even higher negative correlation between reaction time variability and IQ. Columnist Dan Seligman explained this by saying brains with more dense and redundant wiring will get the message consistently, while low IQ people have brains resembling a bad phone connection. Sometimes the message gets through clearly. Other times there’s static.

When I was a teenager (I’m now in my 30s) I discovered that there were many people on the internet who were as fascinated with IQ as I am. However unlike me who has a kind of dark gloomy view of IQ (comparing it to destiny), some of these people were incredibly positive and were big believers that IQ could be improved through practicing chronometric games.

At the time there was a popular chronometric game called Thinkfast that promised to boost your brain power (kind of like the lumosity of its day). At the time I dismissed this as nonsense and for years I avoided even purchasing the game, let alone playing it.

Then I read about a high school project where a bunch of high school seniors played Thinkfast for an hour a day for three weeks, and their highest Thinkfast score was correlated with their SAT score. The results were astonishing:

The correlation between SAT scores and Thinkfast Version 2+ was a potent 0.71. According to Arthur Jensen, the correlation between two mental tests is a product of their factor loadings, so assuming chronometrics and SAT scores share no variance other than g (general intelligence) and assuming the SAT has a g loading of about 0.8, then the 0.71 correlation implies maximum Thinkfast score correlates 0.71/0.8 = 0.89 with general intelligence.

It seemed that ThinkFast was virtually the most accurate measure of general intelligence ever invented. Other intelligence tests (the WAIS) also have g loadings around 0.9, but they were nowhere near as culture fair as Thinkfast. And on an IQ test you can have a bad day, and then you can’t take the test again because you’ve had a practice effect. By contrast with Thinkfast, if you have a bad day, you just try again the next day, and the day after that, and after that…as many times as you want…because it’s your maximum score (your physiological limit) that most correlates with intelligence, not your initial score. Indeed your initial score is culturally biased because some folks have more video game experience than others, but the maximum score one can obtain after practicing an hour a day for weeks is a physiological measure of the brain’s physical limits.

Although I was fearful about how I would perform on an intelligence test I had so little conscious or psychological control over, I found the courage to purchase Thinkfast (version 2+) and invited my pseudo-retarded friend to the house to play with me. The final score was a combination of speed, accuracy, and consistency on six games. The first game measured simple reaction time. Another game measured choice reaction time.

Game 4 (my personal favorite) measured working memory speed. The computer would flash a picture of a common object or scene with a very basic word underneath describing it and you had to press one arrow key if the picture matched the word and a different arrow key if it didn’t. Then it would get more complex (the word REVERSE would appear on screen) which meant that if it did match, you had to press the key which meant it didn’t match, and vice versa, and you would have to do all this with as rapid and consistent speed as possible. The final game measured working memory capacity (parallel processing). A bunch of abstract shapes would flash on the screen and then a few more would flash, and you’d have to press a key indicating from memory whether the new shapes were part of the previous set.

After playing dozens and dozens of times, I maxed out at a score of Brain Master+1 (45 units), and my pseuo-retarded friend maxed out at Theta Gold (41 units). To put that in perspective, among 1400 graduating students at a typical U.S. high school, where all the top talent were playing ThinkFast, only 14 reached Brain Master (43 units) or higher. From here it was estimated that Brain Master (43 units) equaled the 99 percentile in America (roughly IQ 135 on U.S. norms) and I was one whole level (and two whole units) above Brain Master. To put that in perspective, the average doctor, lawyer, PhD, Ivy League student etc, have IQs in the mid 120s when measured on random tests (not the ones used to select them).

And yet at the time I was friends with literally some of the smartest people in the entire World and so I would exaggerate my score by nine levels. Big mistake. Huge! My social IQ was not high enough to realize that lying about your IQ makes about as much sense as lying about your height. Smart people can see the real number whether you admit it or not. Later I would learn that falsification (lie scores) is negatively correlated with IQ, but lying about IQ in particular is especially offensive to high IQ people because it devalues the very trait that makes them special. One incredibly high IQ friend (180+) mentioned that over the years he had met dozens of people who supposedly had IQs above 170. His estimate for the actual IQ of these people? About 115.