<p>The SAT and ACT scores are based off of a bell curve. The middle 80% is given only about 3 times the point range of the top 10% and the bottom 10%, ostensibly to reflect the range of actual intelligence - very few on the top and bottom, with most in the middle. </p>
<p>Unfortunately, many, many people seem to think that the scores are linear, not logarithmic, and that the difference from 700 to 800 is just as significant as the difference from 600 to 700. I've seen it here and I've seen it at school - people retaking when their score is in the lower half of the 700s instead of the upper half of the 700s, as if the one point they missed really reflects badly on their intelligence. People think of the points as a set unit, and that 10 points lost anywhere is the same. Or maybe people really are only satisfied with being in the 99th percentile, but somehow I doubt that - the scores are set up to make the difference appear significant.</p>
<p>So, why not just change the scores so it's linear? Sure, there might be more perfects, but at least then understand the scores would be much more intuitive. The scores would also be a direct representation of your comparison with your peers - instead of being curved to delude students that the difference between a 740 and a 790 is just as large as the difference between a 540 and a 590 (1% vs 10%). </p>
<p>After all, the point of standardized testing is to compare people, right? Why do we have this awful, unintuitive system based off of IQ test scoring (which, btw, the SAT is no longer, and the ACT never was), when it obfuscates your comparison, exaggerating it at the extremes and shrinking it in the middle?</p>
<p>The point is no one viewing the scores as a linear scale but more like a semi-parametric indicator. If you don’t like the way ACT/SAT scores are, you would not like GPA at all.</p>
<p>If by “linear” you mean a uniform distribution on [200, 800], that wouldn’t work since a score of 800 only denotes the top 1/61 or so (i.e. the cutoff for 800 would be much lower).</p>
<p>Or if you are referring to a score that is directly proportional to percentiles (e.g a score of 89 indicates 89%ile), that would also work, but the current 200-800 system conveys the same thing. The score distribution across the many thousands of test takers is roughly normal, so it would make sense to assign a normal distribution (why they use 200, 800 as bounds is another question). The issue is how you perceive the scores.</p>
<p>The scores sole reason for existence is college admissions. And college adcoms know what the scores mean. Heck, CollegeBoard is a non-profit whose members are the colleges themselves. If colleges wanted different test scores — and they don’t — they’d tell CB to change the scale.</p>
<p>@MITer94 Yes, the scale I think is better would be directly proportional to percentiles. </p>
<p>The problem with the current system is perception - it makes it seem like the differences between a 700 and an 800 is bigger than it really is. I think it’s not too unreasonable to assume that part of the reason people, especially here, get so stressed out about their SAT scores (besides the whole determining large chunks of their life…but that’s unavoidable), is that getting a very high score requires such a high percentile. And if colleges don’t care if you’re 97th or 98th or 99th percentile (it seems that even the top top schools only care that your scores are in the 700s, because, after all, it’s really not a huge difference), there’s no reason to have excessively detailed scores at that range.</p>
<p>@bluebayou But the scores aren’t just for the colleges. They’re for the students, too. It is, after all, supposed to tell us how prepared we are for college. Curving it is, IMO, misleading. </p>
<p>Correction, the scores are for the students to send to the colleges for the SOLE purpose of college admissions.</p>
<p>
</p>
<p>Just yours.</p>
<p>
</p>
<p>Actually, the true break point is 750, which is what most unhooked applicants have to the most highly selective colleges. Above 750 the Ivies don’t care about one or two extra questions. </p>
<p>What? How is it subjective that giving the top 5th percentile a point range equal to the next 10th percentile will distort perceptions of the score? A convenient property of numbers is that the difference of 50 between 600 and 650 is the same as the difference of 50 between 650 and 700; the logical assumption is that the number in your SAT score would work the same way. It’s not even a conscious assumption - it’s like the trick grocery stores use when they say a bag of candy is 4.99 instead of 5 dollars. </p>
<p>The placement of the “break point” (or even existence of one) is inconsequential as well. If it’s really 750 instead of 700, then that’s still 1/12 of the scale denoting a tiny, largely irrelevant difference that even colleges don’t care about. </p>
<p>It is just a scale, it does not really matter. Do you know the Richter Scale for Earthquake or the noise level in decibel are in log scale? Do you know temperature units are also arbitrary? The admission offices have this in mind that they know 750 is better than 700. They don’t care if the difference between 750 and 700 is the same as 550 to 600 or not as each schools are just looking for the students within their admission range.</p>
<p>That is your mistaken assumption. Standardized tests go for a bell-shaped curve, with a cap at the top end (which, for the SAT is 800), but a huge tail on the bottom end.<br>
50 points between 500 and 550 is not the same as 50 points between 700 and 750 – and never was, by design.</p>
<p>(Have you taken AP Stats yet? If not, I highly recommend it.)</p>
<p>@JustASomeone this problem of “perception” isn’t their problem; honestly it’s more your problem (or whichever student thinks 6 SD above the mean is significantly better than 4 SD). If we assume that SAT scores on a section are normal with mean 500 and SD 100, a score of 700 places you better than 97% of test takers, and a 750 is at the 99%ile.</p>
<p>@billscho is also right; one might perceive a 7.0 earthquake to be slightly stronger than a 6.0 earthquake, when in fact the shaking amplitude of a 7.0 quake is 10x larger.</p>
<p>Most of these problems are due to one’s own mistaken perception, not the mistake of the CB.</p>