Reuters Investigation: New SAT Design Disadvantages Low-Scoring Students
The story by Reuters uses internal documents to make the case that the College Board's use of wordy math problems made it much harder for some types of students to finish that section of the test than the two English/language arts sections. The text-heavy problems appear to put weaker readers at a disadvantage, the story says.
Reuters quotes from internal emails and other documents that suggest that College Board officials knew of the potential problem with the word-heavy math questions because outside academics raised the issue as they reviewed items while they were being developed. Reuters says that a confidential internal timing study showed that in an early trial in 2014, only about half the students were able to finish the math section, compared to more than 80 percent on the English sections.
The problem was particularly pronounced, Reuters says, among students who typically score low on the test. Low-income, black, and Hispanic students tend to score lower on the SAT. The Reuters report draws on internal emails to claim that the College Board knew of the problem with the text-heavy math questions, and planned to make revisions, but never did.
The report poses a huge potential problem for the College Board only six months after it rolled out its redesigned SAT. A lot is riding on the perceived validity of the SAT, since millions of students submit them to colleges in the admissions process, and, also, since a growing number of states are using the SAT to measure high school achievement, or to report their progress to the public as required by federal law.
When David Coleman, a leading architect of the Common Core State Standards, took over as president of the College Board in 2012, he said he wanted to redesign the SAT to reflect those academic expectations. One of the key concepts in common-core math is that students should master math concepts, and be able to solve math problems in various real-world contexts, not just perform calculations.
Test questions that aim to situate math problems in real-life context could pose challenges for students who aren't strong in reading, or those learning English, something educators have worried about for some time.
College Board spokeswoman Sandra Riley wouldn't provide statistics to Reuters indicating how many students were able to finish the math sections on the new SAT, but she told the agency that the College Board analyzed 10 tests given this spring and found completion rates for English-language learners "are similar to what we see for all students." She also said the completion rate on the math section of the new exam is "equal to or higher than" the completion rate of that section on the old SAT. A College Board spokeswoman declined to add further comment Wednesday in response to an email from EdWeek.
Manuel Alfaro, who was involved in the SAT redesign before being fired from the College Board in 2015, has alleged in a series of posts on LinkedIn that the new test is flawed because of a rushed process that took shortcuts with the item-review process. In one particular post, Alfaro focuses on the question of text-heavy math problems.
The FBI searched Alfaro's home last month as part of an investigation into a breach of security that allowed Reuters to obtain 400 unpublished questions from the SAT, Reuters has reported as part of its investigation into large-scale cheating on the exam. Documents in the case that led to the search of Alfaro's home appear to be under seal, as we reported.
The College Board has said Alfaro, whose title at the College Board was executive director of assessment design and development, can't speak with authority about the development of the SAT, and has noted that he worked there less than two years.
Get High School & Beyond posts delivered to your inbox as soon as they're published. Sign up here. Also, follow @cgewertz for news and analysis of issues that shape adolescents' preparation for work and higher education.