The latest “SAT Understanding Scores” document appears to now be updated for 2017. Compared to the 2016 file, percentiles appear to have gone up (at least for the scores I checked).
https://collegereadiness.collegeboard.org/pdf/understanding-sat-scores.pdf
Let me know if I am interpreting incorrectly. Thanks.
I have not seen the 2016 percentiles. If they have gone up in 2017, does that call into question the 2016 Concordance Tables?
I was able to find a cached version of Understanding Scores 2016. https://webcache.googleusercontent.com/search?q=cache:NY_Bvq2SiPAJ:https://collegereadiness.collegeboard.org/pdf/understanding-sat-scores-2016.pdf+&cd=1&hl=en&ct=clnk&gl=us
The User percentiles are higher for 2017 for at least some scores. According to these documents, the 2016 User percentiles were based on a research study, whereas the 2017 User percentiles were based on the Class of 2017 New SAT scores. Does this indicate anything significant about the quality of the original research study? Do the differences call for new Concordance Tables?
The College Board has removed the document containing the 2016 percentiles from its Web site. However, the document is archived at the Internet Wayback machine, so I retrieved it for comparison.
The percentiles are the same for the Nationally Representative Sample, which are “derived via research study samples of U.S. students in the 11th and 12th grade weighted to represent all U.S. students in those grades, regardless of whether they typically take the SAT.” However, they differ for the SAT User group, which was defined in 2016 as “derived via a research study sample of U.S. college-bound students in the 11th and 12th grades, weighted to represent students who typically take the SAT last as 11th- or 12th-graders” and now defined in 2017 as “the actual scores of students in the graduating class of 2017 who took the new SAT (first offered in March 2016)”.
Here are the differences for the 90th percentile and above:
2016 User group (based on a research study sample):
1550-1600, 99+
1510-1540, 99
1470-1500, 98
1450-1460, 97
1430-1440, 96
1410-1420, 95
1390-1400, 94
1380-1380, 93
1370-1370, 92
1350-1360, 91
1340-1340, 90
2017 User group (based on actual scores):
1530-1600, 99+
1480-1520, 99
1450-1470, 98
1430-1440, 97
1410-1420, 96
1390-1400, 95
1370-1380, 94
1360-1360, 93
1350-1350, 92
1340-1340, 91
1320-1330, 90
In the range of 91st percentile and above, the actual scores were 20-30 points lower than those expected from the 2016 research study.
@evergreen5 I do wonder about the definitions of the groups. The 2016 percentiles mention something about taking the SAT “last” and interpreting it as the “last testing” of the students. However, the 2017 definition is for “students in the graduating class of 2017 who took the new SAT”, with no mention if the College Board used the “last” test or all tests for each graduating student. Usually, the last test is the best one, both because students improve their test-taking over time and because students are more likely to take the test again if they got a lower score than they wanted or expected. If the 2017 definition is over all tests and not each student’s last one, then that could explain why the 2016 scores are higher.
In the old SAT/ACT concordance table before 2015, it also used only senior student scores.
@mdphd92 Thanks for noticing that difference. If anyone else cares to parse the College Board’s descriptions:
2016 percentiles:
2017 percentiles:
@scoodoo1 Any idea how recently this 2017 percentile document appeared on the College Board’s website?
@scoodoo1 I can’t find that second set of numbers anywhere on the college board site- the 2017 group. Is that where you found it?
@homerdog It’s in the gray box about halfway down this page: https://collegereadiness.collegeboard.org/sat/scores/understanding-scores/interpreting, called “SAT: Understanding Scores 2017”
Thx @evergreen5
So what does 99 plus even mean? How much better is it than a 99th percentile?
I guess I feel a little lucky since S19 has another year before he applies to schools and we will have a better idea as to where a 1540 really puts him in comparison to the 2018 kids.
@homerdog
1540 is a top score regardless! Relax and enjoy the journey…
There is no reason to question a 1540 under any circumstances, so probably best to dial back the worry about that.
Thanks all. I’ll take the advice.
Here is the ACT percentile chart:
http://www.act.org/content/dam/act/unsecured/documents/MultipleChoiceStemComposite2017-18.pdf