SAT Writing section score finally counts towards college admissions

After three years of waiting for a clear response from the world of higher education, we are getting increasingly consistent signals that the Writing component of the SAT will count towards college admissions for the majority of our students. Since the initial administration of the Writing section of the SAT in March of 2005, schools have been compiling data on the correlation between performance on the Writing section and performance in college. The data set has grown large enough for robust statistical analysis, and the results are fairly clear: Writing is a significant predictor of collegiate performance. The evidence has been strong enough to move schools from the “let’s wait and see” phase to the unequivocal “Writing counts!” phase.

Does it really come as a surprise that writing ability would be a useful predictor of college performance? The majority of college students will have to turn in hundreds of written assignments during college. The question that is more meaningful is how effectively does the SAT Writing test assess writing ability? What are we actually testing here? Does the ability to identify grammatical errors coupled with the ability to put together a formulaic essay truly reflect one’s writing ability? At this stage we cannot conclude that success on the Writing test connotes writing ability. But the studies conducted to date point to the conclusion that students who do better on this test tend to achieve higher GPAs during their freshman year of college. And that information alone has moved many schools to accept this assessment into the college admissions process.

When and why did this change occur?

In the fall of 2008 analyses began to appear exploring the relationship between the SAT Writing score and academic performance in college. The largest single study was conducted by the College Board.

College Board psychometricians gathered data from over 150,000 students to determine which factors contributed and to what degree they contributed to collegiate performance (measured as Freshman GPA). The study revealed that a student’s high school GPA is the strongest single predictor of college performance. The correlation coefficient measuring the relationship between high school grades and college GPA was 0.36. The SAT by itself came in just behind high school GPA with a correlation coefficient of 0.35. When you combine high school GPA and SAT scores, you end up with a more robust prediction, yielding a correlation coefficient of 0.46.

The College Board researchers also explored the predictive power of individual SAT sections in isolation. If you had to chose only one SAT section to make a prediction of how a student would perform in college, which section would you choose: Math? Critical Reading? Or Writing?

It turns out that Writing takes the cake. It is a better predictor of college performance than either Math or Critical Reading. Here are the correlation coefficients between the SAT sections and College GPA at the conclusion of freshman year:

Writing: 0.32

Critical Reading 0.29

Math: 0.26

Other studies emerged supporting these findings. A notable study emerged in early 2008 authored by a team of three researchers at the University of Georgia: David Mustard, Christopher Cornwell, and Jessica Van Parys, economists at UGA’s Terry College of Business. This study found that of the three sections of the SAT, the Writing section was clearly the best predictor of academic performance in college. The researchers analyzed data from 4,300 test takers and found that with each 100-point increase in SAT writing scores, first-year students pick up 0.07 points on their GPA, 0.18 points on their English class grades, and they earn 0.54 more credit hours.

And is it really so surprising that Writing is the best predictor?

Not to the experts who have been conducting research on the Writing section for much of the last decade. The College Board added Writing to the SAT as a result of a 2001 study which touted the predictive power of the SAT subject tests. In that pivotal study conducted by University of California researchers Saul Geiser and Roger Studley, SAT subject tests (AKA SAT IIs) were shown to rival the SAT in their ability to predict college GPA. Spurred on by these findings, Richard Atkinson, then president of the University of California system, proposed to discard the SAT I reasoning test and instead use the SAT subject tests for admission purposes. But the College Board quickly moved to appease Atkinson and preserve the SAT’s lucrative role in UC admissions. Working with UC officials, the College Board agreed to take the most popular SAT subject test, Writing, which had been around for decades, and incorporate it into the SAT, yielding the “new” SAT. (In an ironic twist, the UC system recently announced that it would cease using the highly predictive SAT subject tests for admissions due to concerns that using these tests might privilege applicants of elevated socio-economic status.)

Well before the 2001 UC study emerged, a cadre of highly selective schools had been using SAT Writing test scores to select their incoming classes. All of the Ivies and dozens of the most selective colleges and universities in the nation have required the SAT subject tests in their admissions process for decades. These schools have accumulated voluminous data validating the predictive power of the SAT Writing test. Schools less familiar with the SAT Writing test, forced to reckon with it only as a result of the 2005 SAT makeover, needed time to gather data and run their own analyses. But that time has effectively passed.

In the fall of 2008 numerous colleges and universities officially announced that they would begin to incorporate the SAT Writing score into their formal admissions process. UGA was quick to come on board after its internal study validated the Writing test. GA Tech was fast to follow suit and adopt the 2400 SAT scale. According to admission reps at GA Tech, Writing would weigh in at the same level as Math and Critical Reading. Dozens and dozens of colleges across the country have made similar announcements: Stanford, UVA, Vanderbilt, University of Alabama, Boston University… the list goes on and on. In time, it seems the majority of schools will come to embrace the SAT Writing test and use it as a tool for admissions.

Why is the SAT Writing section the strongest predictor of Freshman year grades?

If you look closely at the content of the SAT, it is not surprising that Writing scores generate the strongest prediction of college grades.

The Writing test involves error identification, sentence and paragraph improvement and a 25-minute essay. Critical Reading involves vocabulary strength via the sentence completions and reading comprehension. Math involves geometry, algebra I and II, and a smattering of other math concepts. What content would be the most useful to the greatest number of college students?

Let’s first consider the Math test. How many students will actually need to draw upon their Algebra I and Geometry skills once they arrive in college? Aside from engineers, or math and physics majors, few students will need to tap into the treasure trove of middle school and early high school math techniques in their college level classes. These subjects will be useful to a relatively small proportion of college students. Consequently, of the 3 SAT sections, Math is the least predictive of success in college.

Next, we have Critical Reading. This section leans heavily on vocabulary strength and carries the imprint of the early test-makers who assumed one's level of vocabulary to be a metric of innate intelligence. The SAT has been shedding its vocabulary-laden sections with each subsequent iteration of the test. Gone are the antonyms, the synonyms and the analogies of old, but sentence completions remain and continue to emphasize vocabulary strength. In college, knowing a lot of fancy words will impress some teachers, but only if your core fundamentals are solid. In many classes, having a big vocabulary will be of little use. Reading comprehension, one would assume to be very useful in college, but SAT reading comprehension is unique. Many of the question types found on the SAT will never appear in any college English class; they exist only in the rarefied domain of standardized testing. All in all, the Critical Reading section has some merit and is the 2nd best predictor of collegiate success.

That leaves Writing. Why is this section the best predictor of freshman year grades? Simply because most students enrolled in 4-year colleges and universities have to write a lot of papers. Papers that are poorly structured and riddled with grammatical errors will yield lower grades, especially in the liberal arts classes and introductory English classes typical of the first year of college.

Students ill-prepared to write at a college level will naturally struggle in college and may even drop out. To address this, some colleges are using SAT Writing scores to identify at risk-candidates, those who may benefit from remedial writing classes before the start of school or during the first semester. With richer information, colleges can better allocate their resources to serve the students who are most at risk.

These days many students are entering college “at risk” due to significant writing deficits. I’ve heard educators (and employers hiring recent college graduates) voice concerns about the inadequate writing skills of many young people today. If in fact this generation lacks the requisite writing skills to thrive in the world of academia and the workforce, what is the root cause? Do we point our fingers at Facebook, Twitter, or Blog posts, which leave our worst writing habits uncorrected and can even have a reinforcing effect? Are we sacrificing rules of grammar to the god of expediency? Or perhaps our high school curricula have shifted away from grammar and writing skills? I don’t know the answer. But I know from personal experience that quite a few of my students, juniors and seniors in high school, entered our first sessions lacking many of the core fundamentals of writing. This is troubling, but it also presents a real opportunity.

Can we teach students how to Master the Writing Section?

Unequivocally, yes. Of all the sections on the SAT, the Writing section is the most coachable section, by leaps and bounds. I have individually worked with 33 students preparing for the Writing section (and many, many more who did not prepare for Writing). With these 33 students I spent an average of 1.7 hours preparing them specifically for the Writing section (we spent the vast majority of our time working on CR and Math). After 1.7 hours of preparation, my students picked up an average of 72 points in Writing. That’s 42 points per hour of preparation. Of these 33 students, the 12 students who prepped for longer than 2 hours (a mean of 3.2 hours) achieved an average score increase of 106 points. Here are a few examples of the score increases I have seen: Andrew, 650 to 780; Erica, 550 to 710; Lauren, 640 to 750; Mike, 520 to 680; Elliot, 610 to 720; Becca, 620 to 790; Rick, 500 to 620.

My point is this: the Writing section is highly coachable. The fact that Writing now counts is a boon to the test-prep industry. Writing gains are low-hanging fruit; they are the easiest to come by of any section of the SAT.

Why is Writing so coachable?

Teaching someone how to read is no easy task: just ask anyone who has tried. Reading comprehension is a very complicated process involving higher order cognition, metacognition and the like. Critical Reading points, therefore, are the hardest to earn. Math is in the middle. There are a limited number of Algebra I, II and Geometry concepts tested on the SAT. If you master these concepts, your score will improve. Writing, however, involves the most limited set of skills. Of the 16 principal grammar concepts tested on the SAT, 12 concepts appear consistently and account for 92% of the tested content (parallel structure 15%, improper verb tense 13%, subject-verb agreement 11%, pronoun errors 9%, misplaced modifiers 8%, etc.). Master this discrete set of skills and your grammar score can take off. The essay is profoundly formulaic. Once you understand the grading rubric (see my In Praise of Folly: Writing the SAT Essay on this) you can plug in any content into a predictable structure and achieve a respectable score on the essay.

Are we actually turning our students into better writers?

I can help students pick up 100 points on the Writing test, but am I making them better writers? I am making them more aware of the rules of grammar, providing them with helpful mnemonic devices and grammar “mantras” to help them identify and avoid typical mistakes. Once students learn to identify dangling modifiers and comma splices, one could argue that they have become better writers. By teaching students how to create a tightly structured essay, no matter how formulaic, I am giving them a skill, albeit one of greatly limited use. I have my biases, but I feel like my students are leaving our sessions more confident in their writing abilities, and parents have echoed this sentiment. I think, in a modest way, we are giving our students a transferable skill, one that has application beyond the SAT.

What does this mean for you?

If you have a child prepping for the SAT, put some energy and time into the Writing section. In the past, focusing on CR and Math was the smart move, but now, many schools want to see a strong Writing score as well. Do your homework and find out if the schools your child is applying to will use the Writing section for admissions. You can look this information up for many schools on www.collegeboard.com. Enter the school name in the College Quickfinder dialog box, and you can find out how the Writing score is being used under the SAT/ACT/AP tab. Additionally you can determine how individual schools will use the Writing score by going to college web-sites and by contacting admissions departments directly.