The SAT is so bad, it makes you want to Cheat
The recent high-profile sting operation of pay-to-play university admissions, while not entirely unexpected, was certainly shocking in its overt nature. The racial tilt of the education system in America – from a history of school segregation, to the poor quality of teachers and class availability in minority-concentrated schools, to the socioeconomic advantage that yields better test preparation and access to private schools, to the wink-wink donations and legacy admissions – was never in doubt, but it's unlikely that most would have expected the wealthy (and mostly white) to resort to actual bribing and lawbreaking when the existing infrastructure is already so favorable.
Though some of the more entertaining details of the story centered around parents pretending their children played bougie sports like lacrosse or fabergé egg tossing, some even resorting to digitally editing their faces in photos of actual athletes found online, some of the more relatable examples involved the Scholastic Aptitude Test (or SAT) and its bastard brother, the American College Test (or ACT). Parents would – sometimes (incredibly) without the student's knowledge – bribe the proctor to change the answers to the correct ones after the test had been completed, or find a way to get their student diagnosed with a learning disability so they would be eligible for special testing accommodations such as more time. While part of this behavior stems from a belief that cheating has a very low likelihood of getting caught, a non-negligible amount of it probably results from the notion that the SAT and ACT are just dumb obstacles on a student's way to college. I've never heard a student who has taken the SAT said, while taking the SAT, that it's a productive or worthwhile experience. I can't imagine that anyone – apart from some dipshit contrarian professor from Michigan State – would call it a good test, especially in the context of college preparation.
Having taken the SAT three times, I can definitively say that the SAT is a test that you can learn to beat (and while I never took the ACT, I would imagine that it's not all that different in that respect). My first try, I got a 2070 – this was of course when it was scored out of 2400; it's since been changed back to its original scoring system of 1600. My second try resulted in a 2110, and my third (and final) try yielded a 2230. While it's only three data points, it's definitely (logarithmic) improvement, and if you plot the scores over time, it would be pretty linear. And that was without any classes, or having to get any materials. (No, I'm not trying to brag – lots of my friends surpassed the vaunted 2300 threshold without having to take it three times. If I wanted to brag, I'd have brought up my GRE score.)
What does the SAT measure, really? Not intelligence. Not work ethic. Not a cumulative understanding of all the subjects in high school. Income? Maybe that's all it's supposed to measure. And should any test, let alone one of the most important tests you will ever take, really be a 'beatable' test? Shouldn’t it actually measure things that matter, instead of things that, well… don't?
Quick Tangent: When has a student ever used or owned a number one or number three pencil? The lower the number, the darker the marking, so unless you're an artist, chances are that your entire fleet of pencils is number two. I think we're past the point where teachers and proctors have to specify that our pencils have to be number two. God forbid our filled in Scantron bubbles be slightly darker than usual.
Let's be clear about what I mean by beatable – if you know the right things to do, you can almost certainly improve your score each time. If you look online, people have compiled the 100 most common words the SAT will ask or include. There are also formulaic breakdowns of the most common mistakes for each type of question, like parallel structure and 'less' vs. 'fewer' in sentence corrections, or using outside knowledge instead of what's actually in front of you to answer questions about reading passages. This isn't even hidden information, but most people just don't know the right things to look for. Most students incorrectly assume that they just need to do more practice problems, instead of actually understanding why they missed the questions they did. And that's a normal approach, because that's the way it would be if the SAT was actually a good test that didn't follow the 'beatable' mold. If you're taking a calculus class, and you mess up a problem and get the incorrect answer, you probably need to do more practice to sharpen your technique. If you didn’t know the technique, you wouldn't even be able to attempt the problem in the first place. You can't ace a calculus test by just 'knowing' what types of questions are being asked, you actually need to understand the fundamentals and go through a process to determine the approach that will help you get the answer.
Quickly, for the realists – yes, I understand that standardized testing is a profitable venture for test-administering and test-prep companies. None of that will change if the test improves. Students will still have to take the test several times, parents will still panic and buy boatloads of study guides and classes, and the cycle will continue. There's no economic reason why the test can't just be… a better test.
The SAT, as it's currently constituted, has three parts. Reading and math are each scored out of 800 and make up the primary cited score out of 1600. The writing section is a prompted essay that is scored out of eight points each on the criteria of 'reading', 'analysis', and 'writing', for a maximum of 24 points. Here are some real questions that the CollegeBoard thinks are good problems for the SAT. These are real – I had to sign up for a Khan Academy account and everything – you can see how well you'd do yourself, if you'd like. I wouldn't recommend it.
The Math Section
In a political science class, test scores were determined to be 20 times the number of hours, h, the student studied plus 3. Which of the following functions best describes a student's test score depending on the number of hours, h, that the student studied?
This is a fine question, besides the fact that it asks for a function that 'best describes' the situation instead of one that 'exactly describes' the situation, but whatever. Now, anyone who's (recently) taken anything beyond introductory algebra will probably think this is a moron question, but really, it's not bad! Sure, it could (and probably needs to) be harder, but it makes you turn a word problem into an equation (a very important skill). And if you're taking math classes, which most high school students do, this isn't something you should have to study for, no matter how hard the problem is! You should be able to answer this because you, well, understand math.
In a right triangle, one angle measures x°, where sin(x°)=4/5. What is cos(90°-x)?
This is a beautiful question, but unfortunately probably by accident. There are a few ways to do this. If you know the SOHCAHTOA mnemonic for trigonometry and the Pythagorean Theorem, you could construct a right triangle and determine this analytically. Alternatively, you could recognize that the y-coordinate of an angle and the x-coordinate of its complement are the same, because they create congruent triangles. Or, if you happened to know the identity that sin(x°)=cos(90°-x), then you got the reward of sidestepping all of that and saving a few seconds. The problem can be solved if you have the right piece of knowledge, and if you don't, requires you to use some critical thinking that can be traced back to first principles of trigonometry, an understanding of which is essential.
A semicircle LOMN has center at point O and diameter LM. The sector of the circle formed by angle NOM has area 6.25. The radius of the semicircle is 2.5. What is the radian measure of angle LON?
This one, however, is an example of a math question gone wrong. It'll someday be included in the manifesto for my crusade for the banishment of geometry, and is the antithesis of the previous problem. As someone who has had to do a lot of math, for both work and school, this concept has never come in handy. Sure, you could do it, and even if you don't really know how to do it you could logic it out, but that's not really the point. Who cares what the measure of the angle is? It's just the application of a basic formula and some subsequent algebra, but the formula isn't even important.
Which of the following is equivalent to the expression (n^6 * k^14)^5?
The expression (x^2 + h^2)*(x^2-h^2) can be written as (1+m–p)*x^4 – mp, where h, m, and p are constants. What is one possible value of m?
These are equally horrible, but for different reasons. The first one is a multiplication problem that's just a waste of a test question. The second is just a weird thing to ask – not that it's necessarily hard, it's just hardly necessary.
The number of telephone lines in each country in Central America and the Caribbean is shown above in a dot plot. The number of telephone lines in each country has been rounded to the nearest 250 thousand. According to the dot plot (not pictured here), what is the median number of telephone lines (in thousands)?
You might be inclined to think that this isn't bad, because after all, it's just testing whether you understand the fairly simple concept of a median. The problem is that a lot of students are not required to and probably won't take a statistics course in high school. The median is a foundational concept in statistics, but why cherry-pick a topic from a class that most test-takers won't have bothered with? Sure, you might have learned the definition of the word at some point in some other math, but little offbeat oddities aren't the types of things you should be testing for.
The Reading Section
The authors' central claim in the passage "These Are the Days of Lasers in the Jungle" is that: (a) LiDAR’s opponents have prevented the technology from advancing to a point where it might be scientifically useful, favoring traditional methods; (b) Fieldwork and LiDAR are best used in combination when mapping carbon in tropical forests, in order to avoid human error while maintaining accuracy; (c) LiDAR is as important a technology as MRI scanning or the scientific study of the moon with lasers; (d) LiDAR technology is faster, cheaper, and nearly as accurate as traditional field methods for measuring the carbon biomass on Earth.
I didn't bother copying the actual passage here because the passages are mostly dumb and nonsensical, second only to the questions, which are also dumb and nonsensical. Even without knowing anything about the passage, you'd probably be able to guess that the answer is (d). The 'central claim' is just a code phrase for the most general takeaway, which definitely rules out (b) and (c), and you'd imagine that a piece with that title is likely speaking favorably about the technology, ruling out (a). Turns out, (d) actually is the answer, which makes this the epitome of a 'beatable' question. I didn't even have to read the passage! Worse, asking someone to state the central claim of a passage immediately after reading it is just testing for the wrong thing. You don't want someone to read an article or listen to a podcast and only be able to articulate the tweet-length version of it (which in reality, you can probably get from the title or subtitle), you want them to know actual details.
In the first paragraph, the words “disrobed,” “unveiling” and “deconstructed” primarily serve to (a) highlight the negative connotations that laser technology currently has, (b) emphasize the extensive reach of laser technology; (c) demonstrate the inherently unknowable characteristics of objects, even with laser technology; (d) implicitly compare lasers to other forms of technology.
This is from the same lasers passage, and just like before, you don't have to even glance at the passage to make an educated guess. Knowing that it's a pro-laser piece is enough to guess (b), and the words they cite back that right up.
Even if the questions were dogshit, the science passages were at least straightforward and actually somewhat interesting. But don't hold your breath, there are passages that are borderline unreadable, like Louisa May Alcott's 1876 novel "Rose in Bloom." What's the point of being able to read something written in the style of 1800s?! The questions for these fictional works are absolutely horrible – for the Alcott passage, we had the unholy trinity of (1) "over the course of the passage, the attitude of the cousins shifts from…" (2) "which statement best characterizes the relationship among the reunited cousins," and (3) "the conversation between the lady and the stranger serves primarily to…". That's not how people read fiction! Fiction, unlike an informational article, isn't paused every few paragraphs for a detailed analysis and retrospective. That's why English classes are barely any fun. This is the exact type of fundamental misunderstanding that is rampant in the SAT – one that grossly misidentifies a skill that high-schoolers should have, and generates a question that doesn't even properly test for that shitty skill.
Quick Note: In the same Khan Academy practice test that had the 1876 Alcott passage, there was also a letter written by Thomas Jefferson in 1785! Even when schools require students to read Shakespeare, they encourage purchasing the 'No Fear' versions of classic plays like MacBeth, which provides a side-by-side modernized translation of Shakespeare's archaic dialogue. More egregious was that the Jefferson passage had all the same types of questions as the Alcott one, which is the definition of formulaic. At that point, the hardest part of these passages is staying awake through the entire read, which if nothing else (and there really is nothing else) makes for a good test of stamina. If you really want to test for reading ability, shouldn't it be something more timely?
All of this is without going into the grammar section – nitpicking at slight differences in sentence phrasing hardly makes you a better writer. Why learn about punctuation within sentences like semicolons, when you can just use em dashes to show that you're better than the rules of the English language? Why deal with rhetorical styles when using an interrobang puts the interpretation of a sentence in the hands of the reader?! And why bother understanding the rules of capitalization when there's an aCtUaL mEmE tHaT fUcKs iT uP? Correcting small mistakes in non-complex sentences is a job for copy editors and literally no one else.
For anything that's even mildly important, people will find a reason and a way to cheat on it. I knew someone in high school who, despite being very smart, was caught for physically leaning over to peek at someone else's answers during the actual test. But if you give people a fair test, one that gauges things they've both learned and are expected to use in their role as human adults, perhaps they would be less incentivized to resort to extreme measures to bump their scores by a few points. This isn't an unreasonable ask, because we know that tests like this already exist, even at the high school level! You can complain about how the advanced placement (AP) tests have a weird curve and are probably too long, but you'd be hard-pressed to find someone to argue that the content is trivial or unfair. Even the SAT Subject Tests (SATII) kind of get this right – yeah they try to jam too many topics into each test, but you'd still be expected to hit a 700 (out of 800) if you knew what you were doing.
About a year ago, I wrote a piece called 'An Improved High School Curriculum', where (by using my hatred of geometry) I laid out a course schedule for all four years of high school that only included subjects that were actually useful. Some examples of changes included requiring a statistics course, creating two civics courses, and splitting English into things like media literacy and investigative journalism. Perhaps it's unreasonable to ask that we improve how we test before we improve how we teach. But no one really teaches you how to cheat, and as we've seen, a lot of us are pretty darn good at that.