<p>Actually, there seems to be a bigger flaw in Jason’s (and Mifune’s) reading of the paper as it relates to this person’s question. 2340 is 99+ percentile, and more to the point 1590 is 99.8 percentile. The paper relates admissions by percentiles, not actual scores. In other words they don’t cut it as finely as breaking it down to the first decimal place, which is what would be required to differentiate between 2340 and 2380, for example. One can only conclude from their paper that there is a difference once one crosses from one percentile range to another. Harvard admissions has stated a number of times that once a student is 2300+, they then look to the other academic qualifications of the applicant as well as how they might contribute to the diversity of the incoming class. Believe them, don’t believe them; one’s biases will determine that. But there is no hard evidence to the contrary in that paper. As has been already stated, a person that gets 2380 or 2400 may well have other qualities that puts them over a 2300 candidate, and the study cannot account for that effect despite their statistical analysis.</p>
<p>There is really no big deal if the OP takes the test again, but I personally think they are wasting their time and money.</p>