THIS WEEK
Cover News
Opinion A & E
Sports Intramurals
Calendar Comics
 
YH FEATURES
Exclusive
Archives/Search
Planet of Sound
Speak Your Mind
Pick the Pros
Crossword
 
ONLINE TOOLS
Ground Zero
Sublet Search
Rideboard
Book Shopper
Blue Book Search
 
ABOUT US
the Yale Herald
YH Online
 


SATs cannot prevent discrimination

By Kushal Dave

STRAWBERRY:RED:: (A) peach:ripe (B) leather:brown (C)grass:green (D) orange:round (E) lemon:yellow

See the bias? Of course not. But Educational Testing Services (ETS) provides this question as an example of one flagged by its bias-detecting Differential Item Functioning (DIF) system. Hispanic test-takers, familiar with a "green lemon," apparently perform below the white male norm.

Image
NINA FLETCHER/YH
However, my dictionary makes no mention of green lemons, and I've never heard something being described as "lemon green." While it's a nice idea to write a test requiring no cultural context, any evaluation of linguistic ability also tests, and should test, idiom. After all, this question is also biased against people whose parents, bitter about being allergic to citric acid, never allowed lemons in their home. Similarly, a question with the word "bad" is biased against people who still use '80s slang.

ETS caused a stir this September when word got out about a program called "Strivers." This system would identify test-takers who out-performed those of comparable "backgrounds" with the idea of improving "diversity" in college admissions. These ambiguous terms received no clarification when I called ETS: a spokesperson offered only the promise of a full public report once the research is done.

But why is ETS even bothering? Its web page describes "a careful review process to ensure that [the questions do] not favor —or penalize—any particular group of students." Curious to see how they would make the test fair to illiterate students, I read on. "If the groups actually have different knowledge and skills due to differential educational backgrounds and opportunities, the scores will reflect this." This is a blatant contradiction of the preceding sentence. ETS would probably fail its own tests with this drivel.

ETS should get the hell out of the fairness business and stick to writing questions. Its financial statements show $33 million—or seven percent of its testing revenue—being funneled into its money-losing research effort. As much as I enjoyed being forced to subsidize its pipe dreams in high school, I'd rather see ETS reduce the exorbitant costs of the services it monopolizes through the College Board. Asserting that the SAT could ever be an "unbiased" measure of intelligence is akin to claiming that Christina Aguilera might one day have talent.

The SAT's sole virtue is that it is the one test that all students take. It lets colleges say, "You know, Bob has a 4.0 GPA, but his 800 combined score on the SAT has me concerned." The questions will inevitably favor certain people; the most one can hope is that biased questions cancel one another out. Testing recognition of linguistic subtext is part of separating the wheat from the chaff.

As Ward Connerly, a regent at the University of California, explained this week to a group of Yale students, admitting unqualified people to colleges because of skin color or gender is silly; yet these are the only factors used for the DIF and would inevitably be included in any sort of Strivers normalization. Skin color is neither a virtue nor an indicator of background in our increasingly interracial society.

In a recent New York Times [10/23/99] article sparked by the Strivers controversy, Yale Professor David Gelernter demonstrated why he should stick to computer science. "As students overwhelmingly attest, racial diversity enriches everyone's education," he asserted.

A former president of Harvard University, Derek Bok, was more reasonable. "Pretending that applicants [from under-represented groups] have academic skills they lack does nothing to resolve the real problem, and indeed deflects attention from it," he said.

In any case, such disagreements about definitions of merit make one thing certain. ETS should stop wasting our money trying to determine a student's background and its pertinence to test scores—undermining its own objective scale on that student's behalf. ETS is in the business of numbers, and only qualitative factors can really reflect background. Leave that evaluation to the admissions offices. When the SAT gives us lemons, we should be prepared to make lemonade.

Kushal Dave is a sophomore in Pierson.

Back to Opinion...

 

 


All materials © 1999 The Yale Herald, Inc., and its staff.
Got any questions, comments, or advice? Email the online editors at
online@yaleherald.com.
Like to join us?