Failed Test

Tweaks won't fix the deeper problems with the SAT


The reactions are rolling in this week as 330,000 of America's college-bound teens recover from the newly revamped SAT, the preeminent college entrance examination, and the result seems to be about 330,000 variations on "Tests suck."

"At the end, I had blurred vision," commented one test-taker to SAT researchers looking for some feedback.

"I was really hungry," offered another.

Astute assessments aside, the test rolled out on Saturday was supposed to restore legitimacy to the much-maligned SAT, which has been called racist, sexist and classist for as long as today's test takers have been alive. Universities that require the test open themselves to charges of racism; Universities that don't are accused of reverse discrimination.

This seems a fitting legacy for Carl Brigham, a bona-fide racist who designed the SAT in 1925. Brigham's book A Study of American Intelligence claimed tests like the one he would design proved the superiority of "nordic stock."

While the racism issues linger, it's the wider socioeconomic issue that has preoccupied the tests' makers at the Educational Testing Service (ETS). The conventional wisdom is that the test is just another leg up for rich kids who can shell out $1000 for a test prep course. To some, the likes of Kaplan and Princeton Review have turned good SAT scores into a commodity, another saleable ticket into America's Ivy League aristocracy.

The irony here, best expounded on at length in Malcolm Gladwell's 2001 New Yorker article, is that the grandfather of test prep considered himself a foot soldier of meritocracy. The original SAT was supposed to select the true pedigrees from the herd of workhorses, to separate the sharp kids from the ones who stayed up late memorizing geometry theorems. Stanley Kaplan didn't just want to trump the test. He wanted to transform it into another way for America's hardworking kids to catch up to the lucky ones. This is still part of Kaplan's corporate mythology—helper of the eager beaver, bane of the naturally gifted.

Today the situation is a little less Robin Hood and a little more Robber Baron. Kaplan proved beyond a doubt that the test can be taught; his successors put the price of a class out of reach. A good SAT score might mean you're smart, or it might mean you've got pushy parents with a lot of cash. To try to resurrect some notion that the SAT is actually a yardstick of academic potential, ETS has changed up the test, replacing the analogies with an essay and jacking up the total score from 1600 to 2400.

The claim that a new test will somehow outsmart Kaplan is naive. If the SAT can train 3,000 scorers to judge essays with something resembling consistent criteria, Kaplan can train tens of thousands of college-bound teens to reproduce those criteria. Having spent many a Saturday morning teaching analogies to 20-odd lily-white 16-year-olds, I can say with confidence that anyone can learn how to best a standardized test question. Kaplan and its competitors will only benefit from the need to replace used test guides.

Is Kaplan evil? No, Kaplan is inevitable. Parents are always going to use the resources they have to push their kids ahead. The problem is the SAT itself, or more specifically, the idea that ETS can measure aptitude through a series of multiple choice questions or a five paragraph essay.

Critical thinking is every bit as learnable a skill as multiplication. As anyone who has seen the ravages of a truly horrible education system knows, problem solving takes coaching. Kids with a good education already bring years of this coaching with them on SAT day. The ETS can't suddenly slice through the years of privilege and quantify aptitude in four hours on a Saturday morning.

ETS acknowledges this, which is why SAT, which used to stand for Scholastic Aptitude Test, doesn't stand for anything anymore. ETS instead claims that the SAT acts as a predictor of first-year college performance, which is only marginally accurate. The company's own studies show the SAT to predict a paltry 16 percent of the variance in first-year grades.

There is an understandable drive to impose uniform standards on America's motley collection of educational institutions, but there is no reliable standard by which to abide. That's a good thing, and it's reflective of a society that rewards individual talents rather than rubber-stamped clones.

Fans of the standardized test–happy No Child Left Behind Act, if there still are any, might take a lesson from the long, dirty history of the SAT and drop the urge to impose useless tests on innocent kids.