<p>"Is it that statistics lie or that people don't understand what they mean? The article you mentioned is talking about means, NOT medians. If the 50% SAT range, is 1400-1580. then obviously 25% are above 1580 and 25% are below 1400. It is a very skewed population and the mean and the median are going to differ by quite a bit. You got your 1490 "median" number by taking the mean of 1400 and 1580, but that is extremely unlikely to give you the correct median. It would imply that 25% are from 1200-1400, 25% are from 1400-1490, 25% are from 1490-1580 and 25% are 1580+. In a population this skewed, the median is much more likely to be around 1540 or 1550."</p>
<p>Yes, the Princeton article used the word "average." Yes, average is supposed to be the mean. But do you really think that the mean would produce all nice round numbers like 720, 730, 720. Not very likely. I'm pretty sure those are medians.</p>
<p>As for your point about 1490 not being the real H median. Yes, the distribution is probably not going to break down quite so neatly to give precisely the 1490 figure. </p>
<p>But your figure of 1540 or 1550 (it's now gone up, no less) is just pulled out of the air. </p>
<p>In fact, the differencees between the medians and the means are not as dramtatic as you imagine. </p>
<p>Swarthmore, for example, includes both on its Common Data Set for the Class of 2009.</p>
<p>The middle 50% numbers (median) are 680-770V, 670-760M (1350-1530).</p>
<p>The actual medians: V730, M710 (1440)
The actual means: V719, M708 (1427)</p>
<p>You'll notice that the 1440 figure equates pretty well with the hypothesized 1490 for H. Yes, the Swarthmore numbers are lower than the H numbers. But the difference between the 25% and the 75% is the same 180 points. Same skew, just at a slightly lower (but still high) level. </p>
<p>So until you have a source for 1530 or 1540 or 1550 being the median SATs at HYP, perhaps you should stop repeating it.</p>