Bragging rights for us Quants
The GRE is a fairly well accepted (or at least widely used) test of verbal reasoning, quantitative reasoning, critical thinking and analytical writing skills. Which type of doctoral students do you think would be smartest according to this measure - philosophers or physicists?
The chart below is by Mark Perry based on 2002 test scores of Ph.D candidates in the US (click on it to see a clearer version). Statisticians are not separately listed so I propose that we classify ourselves as (applied) mathematicians for the purpose of this exercise!
Out of the top ten, nine are highly mathematical disciplines. Some have suggested that the GRE tends to advantage quants over poets. But notice that the physicists did relatively well on the verbal while the English students did the best on verbal, suggesting that this does indeed measure roughly the right dimension.
I could not find the details of how the students were selected for test. But regardless, several obvious biases/explanations occur. Physicists are ranked number 1, but I suspectthat under-graduate physics students would not be ranked highly. I recall at La Trobe, some 15 years back, that the enter score for Physics equated to an average fail. The course was simply not popular. So you have to think about the selection process from under-graduate to post-graduate. I tend to think that there are not that many jobs for physicists apart from academia. So the best and brightest continue to post-graduate aiming to be academics. Whereas, in architecture and business the best and brightest may well head off down town after their under-graduate degree.
Public administration is a real worry - the students are actually below average. (The mean Verbal score is 462 and the mean Quant score is 584, though what was the cohort?). And communication students have lower verbal scores than physicists! There goes the stereo-type of the coffee stained introvert.
Notice how medicine is ranked second last! Remember, this is US data where entry to medicine is easier than here and decided after students graduate . Whereas in Australia, the places are deliberately limited by the AMA and based on a quite brutal competition between hyper-motivated 17 year olds from private schools. The result is that Australia probably has the smartest doctors in the world. However, I wodner how many have a real calling to medicine and how many could be better suited to physics than handing out anti-biotics.
As Bruce has pointed out below, there are correlations between quant and analytic (0.78) and so there is a sense in which quant skills are double counted - though there is a correlation of 0.51 between verbal and analytic. The raw data are in Excel format HERE. I wonder whether a factor analysis would be useful here. If both factors had positive loadings then you might just take the average of the two factors for ranking. Unfortunately, the R-function “factanal” will not extra two factors from three variables so I cannot pursue this. If you replace the quant scores by the average of quant and analytic and then rank by total the league ladder looks fairly similar, but philosophy and English are promoted:
PHYSICS, MATH, PHILOSOPHY, ECONOMICS, CHEMICAL.ENG, MATERIAL.SCIENCE, ELEC.ENG. MECH.ENG, ENGLISH, HISTORY, RELIGION.THEOLOGY, POL.SCIENCE, ART.HISTORY, ANTHRO.ARCHAEOL, CIVIL.ENG. INDUST.ENG, ARCHITECTURE, SOCIOLOGY, BUSINESS, PSYCHOLOGY, MEDICINE, COMMUNICATION, EDUCATION, PUBLIC.ADMIN.
You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.
November 19th, 2008 at 2:30 pm
The most worrying data point is for education, for it is today’s educators who train tomorrow’s talent. Arguably, the brain drain from education when women achieved equal opportunity in the states (instead of being railroaded into education) has led directly to the decline in US achievement over the last 50 years.
It’s one case where “throwing money at the problem” could indeed make a difference. If today’s best and brightest saw the most lucrative careers in education, imagine what tomorrow’s generation would produce.
November 19th, 2008 at 3:33 pm
Poets are probably not so obsessed by rankings as quants. If they were they might notice that quantitative and analytical scores seem to be biased up relative to verbal scores, and dismiss this ranking as dodgy.
November 19th, 2008 at 5:11 pm
Shaman. The “education” students listed are actually doing a docotorate in education. I get the impression you are talking about throwing money at teachers. Ph.D’s In education tend to be educational theorists rather than teachers.
John. The fact that the quant scores are high actually means that the quant tests are easier! More to the point, it is not fair to just add the three scores without somehow standising their scale.
November 19th, 2008 at 5:27 pm
First, it seems these test scores are for applicants, not admitted students.
Second, the main test for med school admission in the US is the MCAT. The ‘medical’ students here are probably applying to masters programs designed mainly for failed med school applicants who would like to get another year of education under their belt before another try at Med School admission, so its not too surprising that their scores are low.
November 27th, 2008 at 9:21 pm
It seems to me that the “Quantitative” and “Analytical” scores are reasonably correlated. Perhaps these are measuring something similar and so there is effectively double counting resulting in a bias in favour of the mathematical disciplines.
November 27th, 2008 at 10:10 pm
Bruce, You are right. I managed to get the data into a spreadhsheet HERE. The correlation is 0.785 and the correlation with the verbal scores is close to zero. It sure looks like double counting. These data cry out for a factor analysis. I will do it tomorrow and post the results at the bottom of the post.
December 2nd, 2008 at 2:21 am
On the whole though no huge differences really between them?
Katie
September 30th, 2009 at 10:34 pm
Did the factor analysis ever get done and if so what was the result please?
Sally