## Monday, July 18, 2011

### wired the wrong way

When it comes to the GRE, it seems the effort needed to reach the 99th percentile for Verbal isn't nearly as great as the effort needed to reach the 99th percentile for Quantitative. Why is it that, if you miss two problems out of 60 on the Quant, you're bumped down to the 92nd percentile, whereas the same error rate on the Verbal nets you a 99th percentile? Two possible explanations occur to me:

1. It's actually harder for most people to score 750 or higher on the Verbal. This is difficult to believe, given how I was able to score 750 twice with only a week's practice. If we go with this hypothesis, however, it leads to the notion that most people taking the GRE are math-oriented, i.e., they're not so good with the "fuzzy" logic of the Verbal section, but they generally rock and roll on the math. If so, that's cause for pride: I'm one of the few and the proud when it comes to the GRE Verbal section. But it's also cause for consternation: am I really that dumb when it comes to math?

2. It's not that the Verbal section is harder for most GRE test-takers: it's that the ETS standards are higher when rating the Quant section. This possibility is bizarre, but worth chewing over. The Quant section requires only three kinds of knowledge: Algebra 1, Geometry, and something called "data interpretation," which amounts to using algebra while staring at and interpreting graphs and pie charts.

That's not much math knowledge, when you think about it: by the time most kids are taking the SAT in high school, they've gotten at least as far as pre-calculus (called "functions and analytical geometry" where I went to school), and many have gone further into either AB- or BC-level calculus. Meanwhile, the level of reading skill demanded for the SAT (and, by extension, the GRE) roughly corresponds to the level required to get through the sort of material one might encounter as a high school senior or early undergrad.

If I'm right, then a difference in standards between Verbal and Quant is a distinct possibility. To put the matter more plainly: the SAT and the GRE test higher-level verbal skills while simultaneously testing lower-level math skills. It's not so much that the test-takers are more mathematically inclined as that there's a tacit expectation that a random sample of the population will do better at the math. (An expectation that has, I assume, been borne out by decades of testing data.)

By that reckoning, possibilities (1) and (2) are interrelated, because if expectations are high that the general population will do well at lower-level math, then it doesn't say much about my math skills if I can't break the 92nd percentile (cf. the "cause for consternation" referred to in [1] above).

The inescapable conclusion, then, is that I absolutely have to bone up on my math, possibly even running myself through an Algebra 1 and Geometry textbook just to make sure I've got the basics down. It's a shame my brain wasn't wired better for math. This is going to be one of those cases where effort will have to make up for talent.

_

#### 1 comment:

Anonymous said...

These things are norm-referenced against the population that actually take the tests. That means it's not the general population, it's a rather select group of people who have an undegraduate degree and who want to go on to grad school--a distinct minority. And if the math is limited to basic stuff, that means that a lot of that select group gets a lot of it, so the competition is intense, and one or two errors will count for a lot. That's my theory, anyway.