Skip to main content

Reply to "The Wide Variations in the Black-White Higher Education Gap in America's Largest Cities"

A Strange But Common-Sense Strategy to Lower the SAT Racial Scoring Gap

Black students score lower than whites on the more difficult SAT questions. But the racial scoring gap is smaller on the hard questions than it is on the easier SAT questions. As a result, one educational psychologist suggests that the SAT only ask the hardest questions. The author says that if this revision were adopted, the percentage of blacks in the top-scoring group on the SAT would triple.

The very large racial gap in scores on the Scholastic Assessment Test (SAT) continues to be a major barrier to increasing the number of black students enrolled in higher education. The gap, which stands at about 200 points on the standard SAT scale of 400-1600, or about 15 percent, has been persistent and has even increased slightly over the past decade.

Roy O. Freedle, a retired senior research psychologist at the Educational Testing Service, the company that devises the test questions and administers the testing program for The College Board, has a simple solution. His scoring system, if adopted, would significantly close the racial scoring gap on the SAT. Remarkably, Freedle's solution is to throw out the easy questions on the SAT and retain only the most difficult questions. According to Freedle's analysis, published in the Spring 2003 edition of the Harvard Educational Review, black students do far better relative to whites on the most difficult SAT questions than they do on the easy questions.

ETS regularly assigns a difficulty factor to each question it uses on the SAT. The easiest questions are rated 1 and the most difficult questions are rated 5. However, under the SAT scoring system easy questions count just as much as difficult questions. Freedle's analysis found that if only the 40 hardest questions on the SAT are used to determine the final score, the racial scoring gap would be reduced by at least one third. Freedle's data shows that at almost all scoring levels black students do better compared to whites on the harder questions, but perform less well compared to whites on the easier questions.

For example, Freedle's data reveals that for black and white students who score 640 on the verbal portion of the SAT, black test takers got 69.2 percent of the answers correct on the 40 questions rated most difficult. White students who scored an identical 640 on the overall test got only 68 percent of the most difficult questions correctly. This same discrepancy occurs for black and white students at each scoring level from 200 all the way up to 800. In contrast, when only easy questions are considered white students typically get more correct answers than black students. And this also is true at all scoring levels of the SAT.

Reasons Why Blacks Do Better on Hard Questions

Why do black students do better on the hard SAT questions but perform less well on the easiest questions? According to Freedle, the answer is cultural bias.

For example, Freedle points out that the difficult analogy questions on the verbal portion of the SAT use relatively uncommon words such as "anathema, sycophant, or intractable." These words are not used in everyday parlance among high school students but they are frequently used in an academic setting as vocabulary words. For the most part these words have one distinct meaning and cannot be misinterpreted by any cultural bias. Therefore, black and white students who have learned these words in school will be on a level playing field in trying to answer SAT questions that contain words such as these.

But on the other hand, Freedle says, "It is well known that common words often have many more semantic (dictionary) senses than rare words. Many high-frequency analogy words such as 'horse' and 'snake' have many dictionary entries. Various researchers have hypothesized that each cultural group assigns its own meanings to such common words to encapsulate everyday experience in its respective culture. Thus, individuals from various cultures may differ in their definitions of common words. Communities that are purportedly speaking the 'same' language may use the same words to mean different things... Thus, the cultural and lexical ambiguity that African Americans are hypothesized to experience when responding to many easy verbal items offers one promising explanation for why they do worse. Hard verbal items often involve rarely used words that are hypothesized to have fewer potential differences in interpretation across ethnic communities."

"Snake" May Have a Different Meaning to Black and White Test Takers

Let's look at Freedle's examples of easy questions that could be misinterpreted due to cultural bias. A black student in the inner city may hear the word "horse" used far more often in relation to heroin than a white suburban student who is accustomed to seeing horses in rural pastures. Similarly, a "snake" may be a reptile to a white suburban kid while it may primarily mean a deceitful person to a black kid from the city. In this way cultural nuances can affect how a student interprets what appear to be easy SAT questions using words in common everyday use.

Freedle further argues that only hard questions should be considered on the SAT because results demonstrate the highest level where a student can competently perform. This measure should give college admissions officers a better picture of the type of work a student is capable of completing successfully at the college level. Easy questions on the SAT are comparable to material that is learned in the early high school years or before and therefore do not offer a good indicator of the level of work of which a student is capable.

How would revised SAT scoring boost a black student's chance of being admitted to a selective college? Freedle's calculations show that 0.72 percent of all black test takers are among the top-scoring group on the SAT. But using the Freedle revised scoring formula in which only the most difficult questions are counted, blacks become 2.46 percent of the top-scoring group. Thus, under the Freedle scoring system, the group of black students whose academic qualifications are suitable for admission to the top colleges and universities more than triples.

Those who adhere to a strict merit principle for admission to colleges and universities must admit that an SAT with only hard questions would not violate their criteria for a fair system. All students are being judged equally on challenging academic questions and without regard to race.