Cognitive profile

I’m interested in psychometrics. In particular, I’m interested in the study of the different facets of human cognitive ability and how people’s profiles of strengths vary. I’m also interested in how people compensate using their (relative) strengths in some areas for their (relative) weaknesses in others. For those who share my interests in this type of analysis, I’ve included some information about myself below. The information is a mix of data from IQ tests and other cognitive skill tests I’ve taken, real-world information about me, and some speculative introspection by me that contextualizes things. Please note that any individual test yields only limited information, and the purpose here is not so much to provide definitive information about myself as it is to provide yet another data point to people interested in getting an understanding of how cognitive profiles might work.

Raven IQ test

I have taken the Raven’s Standard Progressive Matrices test at this site (on December 19, 2013). The site reported my IQ as 137. Note that the site uses Standard Progressive Matrices rather than Advanced Progressive Matrices, so measurements in the 130+ range are not reliable (my subjective impression was that of the 60 questions, I had to think carefully for only 5-10 of them). You can access the report here. The test is normed so that scores are on a normal distribution with a mean of 100 and standard deviation of 15 for the United Kingdom population. A score of 137 puts me at roughly the 99.3th percentile of the United Kingdom population.

I also took the Raven’s IQ test at the IQ comparison site on March 29, 2010. This used Advanced Progressive Matrices, and was more attuned to differences at the high end of the spectrum. That test gave me a score of 138 (putting me at the 99.4th percentile of the United Kingdom population). Unfortunately, I do not have a record of that test score, and I am not able to remember my password or the answer to my secret question, hence I am unable to access my score details.

My guesstimate based on this is that on measures closely related to the g factor, my IQ would be in the ballpark of that measured by the Raven’s tests. However, as suggested by Spearman’s law of ability differentiation (also alluded to in my score report), I would have greater cognitive strengths in some areas beyond what the g factor would predict. These are discussed below.

GRE scores

I took the general GRE in October 2006 and received a score of 800 on the quantitative and 690 on the verbal section. In practice tests I had taken, I had scored 800 on the quantitative in all practice tests, and 710, 740, and 800 respectively on the verbal. I believe that getting a question wrong early on in the verbal section of the actual general GRE caused my score to be lower than it would be in the median case.

According to the IQ comparison site conversion tables, my verbal + quantitative score on the actual test would correspond to an IQ of 150.11 using the norming above. Note, however, that for a number of reasons discussed below, I would expect my GRE scores to overestimate my IQ.

Reaction time tests

I have attempted the reaction time test at I had done this test in 2009 or 2010 but did not record my scores. I attempted the test again on December 9, 2013. My first average-of-five-tries was 328 ms, my second was 310 ms, and after some practice, I was able to bring it down to 300 ms. With an external mouse, my average came out a little lower (about 285 ms). These scores are considerably worse than the average of attempts on the site. Mental chronometry suggests that high reaction times are correlated with low IQs, thereby pointing in the direction of my having low IQ.

However, it needs to be borne in mind that simple reaction time tests measure the sum total of stimulus-processing time and movement time. In order to isolate stimulus-processing time from movement time, one would need to vary the complexity of the stimulus and then use a model such as Hick’s law to calculate the slope (the rate of stimulus-process) and intercept (the movement time). I have only taken the reaction time test above. This uses a single stimulus, so I’m unable to provide separate estimates of my stimulus-processing time and my movement time.

Verbal versus visuo-spatial skill

My cognitive strengths lie more in the direction of verbal than visuo-spatial skills. This is based on a self-assessment and real-world performance, not based on cognitive testing. I often convert problems that putatively rely on visuo-spatial skills to corresponding verbal problems. Two examples:

  • When attempting the more difficult Raven’s test problems, particularly the ones that involve patterns along diagonals, I have trouble “seeing” the answer. I know that there are some people who can literally “see” the answer to these problems just as easily as I can “see” the answer to the easy problems that involve vertical or horizontal patterns. I overcome this inability by explicitly counting the numbers and naming the shapes, and then seeing patterns in the counts or names. (Note that calling this an “inability” may be a bit distortionary — most people can’t get these problems either visuo-spatially or verbally).
  • While preparing for Olympiads, I acquired considerable passion and skill for the geometry problems in Olympiads. However, my geometric intuition was built mainly on remembering large numbers of definitions and concepts rather than on skills at visuo-spatial manipulation. I have met people who can manipulate three-dimensional figures in their minds or mentally carry out complicated geometric transformations (such as inversion about a circle). I’m not able to do either — instead, I rely on an explicit knowledge base to help reconstruct what would happen.

Short-term memory

I seem to have significantly better short-term memory than the median world resident, but I haven’t tried to measure just how much better. According to this visual forward digit span test, I am able to do 10 digits with a near-zero error rate, 12 digits with a 70%+ success rate, and 13 digits with a success rate of about 50%. For 14 digits or higher, my success rate is well below 50%. Note that this is without significant practice. Memory techniques would probably help me go higher, but low transfer of learning suggests that this is unlikely to be worthwhile.

Crystallized intelligence

I expect measures of my crystallized intelligence to yield higher values (in relative ranking terms) than fluid measures of intelligence. Reasons:

  • I have considerable obsession with structure and organization of material and ideas. This is closely related to my trait of high conscientiousness (see more about my personality at my personality profile). These traits help me acquire considerable amounts of knowledge and in-depth understanding.
  • I have strong verbal skills, and these are more helpful for storing knowledge and structure in the long term because most ideas of sufficient complexity are stored verbally rather than pictorially.

The following are some indicators in favor of the above hypothesis:

  • My general GRE scores somewhat overpredicting my IQ relative to Raven’s (however, this is confounded by other factors such as the significant role of good quantitative scores, given that I specialize in mathematics).
  • My good performance in Olympiads: I represented India twice at the International Mathematical Olympiad and received silver medals both times. Note that getting to the Olympiad from India is easier than getting to the Olympiad from a top-performing country such as the US or China, and might be considered comparable to getting through the USAMO.
  • I performed well in the IIT-JEE despite not being focused on preparation for that test, since I was not planning to join the IITs. I received a rank of 10 on the screening test (a multiple choice test) and a rank of 158 on the main test (a long-form answer test). This is one of India’s most competitive entrance examinations taken every year by a few hundred thousand students ending high school, with about 10,000-20,000 people who seriously prepare specifically for the test.
  • I took the mathematics subject GRE in November 2006, and got a raw score of 62/66 and a scaled score of 880/990 (97th percentile among test takers). People taking this test are students applying to graduate school in mathematics. In practice tests I’d taken, I’d done marginally better, so I believe that the actual score is probably a slightly pessimistic estimate of my performance.

Basic information