Is the LSAT an IQ test?
Last June, along with thousands of undergrads around the world, I sat for the Law School Admission Test. Law school hopefuls come from many walks of life; some are psychology majors, some are music majors, some hope to fight for the rights of the disenfranchised, some want to work on Wall Street. But in order to get there, each and every budding law student must first take the LSAT.
The LSAT is composed of five 35-minute multiple choice sections: analytical reasoning (logic games), reading comprehension, two sections of logical reasoning, and one random experimental section. These 5 sections are followed by a 30 minute unscored writing sample section. LSAT scores form a bell curve around 150 with a standard deviation of around 10 and a range of 120–180.
Below I’ve included a sample question from a logic games section. Try it out! The answer is at the end of the post.
Given enough time, the games can be brute forced and thus solved accurately by most people. The challenge for most is being able to get through each section accurately within the 35 minutes allotted.
The test is intended to evaluate a student’s reasoning, logical, and comprehension skills and thus looks very similar to tests of verbal IQ or fluid reasoning. It’s important to note that compared to measures of IQ like Raven’s Matrices, there is likely an aspect of cultural bias since the test is in English and familiarity with Western subject matter can aid with questions that draw on real-life (Western) situations.
Regardless, many think that the LSAT and IQ measure very similar constructs and believe that there is a very strong relationship between the two.
Mensa America will accept a 95th percentile LSAT (167), or a Stanford-Binet of 132. The poetic genius society will accept those who score a 174 or a Stanford-Binet of 141. There are several other high-IQ societies and organizations that accept high LSATs as equivalents to standard IQ tests. Are they justified in assuming that there is a reasonable equivalence between the tests?
The Law School Admission Council, which administers the test, only makes the claim that the LSAT can predict a student’s ability to succeed in law school (measured by law school GPA). The correlation has been measured at about r=0.36. That may not seem like a lot, but even undergraduate GPA is less predictive with its correlation of r=0.28.
Personally, I doubt the correlation between IQ and the LSAT is strong enough to warrant percentile-equivalent acceptance. While no formal studies have been conducted on the subject, my experience and those of my peers indicate that the LSAT is far more variable and learnable than an IQ test. While a naturally strong LSAT taker is likely to start out scoring in the 160s (more than one SD above mean), through practice they should be able to pull themselves up to the 170s (something very few people manage without a little familiarity with the test). Someone who starts in the low 150s may have to study and practice four times as hard to make their way up to the 170s, but it happens. These changes are much more drastic than improvements seen on standard IQ tests. Additionally, in the upper ranges, even a single multiple choice question can result in a 2 point score difference on the LSAT, which leaves the relevant part of the scale incredibly susceptible to error.
A recent study from the neuroscience institute at Berkeley found that compared to controls, people studying for the LSAT showed significant decreases in radial diffusivity in white matter connecting frontal cortices and in mean diffusivity within the frontal/parietal lobe white matter, with larger LSAT score gains being associated with greater decreases in certain areas. According to the researchers, these changes are associated with improving reasoning ability and spatial cognition and as a consequence, IQ scores.
While the relationship between IQ and LSAT may not be as strong as some believe, it is undeniable that some of the constructs involved and even some of the neurological factors that contribute to the scores are very similar.
Logic Game Answer: B
References:
Berkeley study:
http://journal.frontiersin.org/Journal/10.3389/fnana.2012.00032/abstract
LSAC study:
http://www.lsac.org/docs/default-source/research-(lsac-resources)/tr-11-02.pdf
http://www.poeticgenius.com/qualifications.htm
-260522465
