I see average ACT scores range for Duke is from 30 to 33 on their website. However, I see posts from people got accepted by Duke most above 34 in Confidential Discussion, so where is this official score range coming from? What is true average accepted scores?
Most schools publish something called the Common Data Set that gives the score range. Google it for Duke. That is probably the 25% - 75% range. Know that a lot of applicants below the 50% mark (so below 32) have some kind of hook – athlete, URM, legacy, or something really compelling about their application. Remember that CC is not a random or statistically valid sample pool.
The current freshman class had accepted 25%-75% ranges of 31-35 for Trinity Arts & Sciences and 33-35 for Pratt Engineering. Of course, the enrolled class is lower, according to the Common Data Set, with an ACT range of 31-34.
http://admissions.duke.edu/images/uploads/process/DukeClass2019Profile.pdf
http://ir.provost.duke.edu/facts/cds/Duke%20CDS_2013-2014.pdf
25% are above that range right? And it is those who are on this site!
The actual number of accepted students who post their stats is relatively small. Yes, it is possible that many that post are in that 25 percentile range that are above the average 75 percentile for the school. This website seems to have a skewed population of posting students.
OP: remember, there are MANY who are rejected who are solidly in or even above this range. Scores are only one aspect. Duke has and will continue to reject kids who get 33s and above.
Note that the majority of applicants in the CC sticky thread have a 34+ ACT, so it is expected for the majority of accepted persons in the CC thread to have a 34+, even if acceptance decisions did not consider score. The acceptance rate in the CC sticky for 34+ ACT posters was ~80%, while the acceptance rate for 32-33 ACT posters was ~70%. That’s not a huge difference, so it does not suggest than one needs a 34+. It also shows that persons who post in the CC sticky thread are not a good reflection of the typical applicant for a variety reasons, so their average ACT score is not expected to align with the ones Duke reports.
It’s also worth noting that the set of students accepted at Duke includes within it the subset of students who enroll at Duke, but is larger and has different characteristics than the “enrolled” subset. (However, 99% of the students accepted ED in December are expected to enroll as well, so that difference won’t be quite so apparent now.) The test score ranges you see on Common Data Sets and various information sites are based on the subset of enrolled students, not the set of accepted students.
In recent years, Duke has accepted about 3,500 applicants per year to fill a class slightly larger than 1,700 – so there are roughly twice as many students in the accepted set as there are in the enrolled subset. About 800 students are accepted ED, so that means the ratio of accepted to enrolled students in the spring RD round is roughly 2,700:900, or 3-to-1.
In general, students with higher scores in the regular round tend to have more options than students with lower scores. It’s likely that the acceptance ratio for students with the very lowest scores is close to 1-to-1, which means that the ratio for students with the very highest scores could be something like 6-to-1. In other words, the proportion of students with high scores in the set of accepted students is likely to be higher than the proportion of students with high scores in the enrolled class, from which most of the published score ranges are derived. In theory, the median score for enrolled students could be close to the 25th percentile level for accepted students, and the 75th percentile for enrolled students closer to the median for accepted students.
A few colleges do publish the score ranges of their accepted students as well as their enrolled students. The figures from those colleges confirm that the effect described above exists, but probably not to the extreme extent suggested. If you want to get a sense of the score range for the accepted class, however, it’s reasonable to add 20 points to the endpoints of the 75-25 range based on enrolled students, especially on the high end…
Ranges (such as 25th-75th %ile) are much more useful than knowing an average or median score. An average lumps all scores together and divides by the number of scores- a few outliers can mess with the bulk of student scores. Likewise a median only tells you one half are above/below that number- it does not tell you what the majority scores are. The middle half tells you that one of the students are here, and you know one fourth have better scores. And one fourth lower as well. Compare it to your kid’s score to see if s/he fits somewhere in the middle or how far it is from that.