Any others? Isn’t Feb. 15 a day when a lot of early decisions are released?
See also UGA.
Someone may have already posted this link (I’m late to the convo & didn’t read all 22 pages closely) – but I found this interesting: http://www.chyten.com/blog/insider-secrets-about-the-october-1-2016-sat-scores/
Based on how my son did on the new SAT, I think the concordance tables are off. He didn’t do well on the ACT & vastly preferred the SAT format, so we went with the new test, even though his 1390 “new” score is like an old 1330, according to the tables. However, his 1390 was from one sitting in Oct (he outperformed his March 2016 test), so no superscore for him. He improved his scores on an apparently much tougher test if chyten’s blog is to be believed. Based on the academic scholarships he’s getting with the 1390, I have to think some schools are not going strictly by the concordance tables, but are adjusting the scores on the fly as data from applicants comes in.
This year, colleges may choose to do one of the following:
- Treat old and new SAT as separate pools. (just like ACT and SAT were treated)
- Convert old SAT scores to New SAT scores. (esp. if majority of the applicants had submitted new SAT. Also it will be easy to compile CDS)
- Convert new SAT scores to old SAT scores. (but why?)
- Just treat new and the old to be the same since colleges are looking for a range anyways, especially 'holistic' colleges.
The problem with concordance tables (https://collegereadiness.collegeboard.org/pdf/higher-ed-brief-sat-concordance.pdf) is that you will get different results depending on which tables you use.
For e.g. take someone with the following new SAT scores.
740 (RW) + 790(M) = 1540
M 39.5
R 38
W 37
1: Using tables 3, 4, and 5 (Most colleges care about individual section scores rather than the total)
780 + 720 + 710 = 2210
2: using tables 3 and 6:
780 + 1450 = 2230
3: Using Table 1 (But remember 600(RW)+ 800(M) is not the same as 800(RW) +600(M) )
1540 → 2260
So, 2210=2230=2260 according to college board’s new math
Or colleges may choose to completely ignore the tables and simply treat new and old to be the same.
E.g. 39.5 20 + 3820 + 37 * 20 = 790 + 760 + 740 = 2290
So, the translation could be anywhere between 2210 to 2290 and that is a big range.
We all know college admission is not a science, but this time all these math, with all the confusions, will become a history soon.
In the end, students will settle in some college, fair or not, and be happy
Vandy ED 2021 (https://admissions.vanderbilt.edu/vandybloggers/2017/02/class-of-2021-early-decision-summary-statistics/)
Old SAT*
Middle 50% Critical Reading: 710-790
Middle 50% Math: 750-800
New SAT*
Middle 50% Evidence Based Reading and Writing: 710-760
Middle 50% Math: 730-790
From UVADeanJ (http://talk.qa.collegeconfidential.com/university-virginia/1934886-university-of-virginia-class-of-2021-ea-discussion-thread-p42.html)
“In the years after a major tweak in testing (1995/1996, 2005/2006, 2016/2017), I would argue that admission officers actually lean on them less. Though the College Board put concordance charts together, the fact is that we don’t have any historical data about the new iteration.”
University of Georgia EA 2021 (http://ugaadmissions.blogspot.com/2016/11/early-action-decisions-are-available.html)
"Mid 50% Admitted Average SAT (old, CR+M): 1300-1480
Mid 50% Admitted Average SAT (new, EBRW+M): 1300-1430"
Fordham EA 2021 (https://fordhamram.com/2017/01/25/early-action-applications-break-records/)
"This year’s total applicants, as of Jan. 18, scored an average high score of 1299 out of 1600 on the SAT. This score is just barely higher than last year’s average high score of 1297.
The average high score of students who are already admitted to Fordham for the Fall of 2017 is 1392. This is just slightly lower than last year’s admitted student’s score (1394)."
Boston College EA 2021 (http://bcheights.com/2017/01/18/2900-admitted-class-2021-acceptance-rate-slightly-33-percent/)
“Admitted students averaged a 33 on the ACT and a 1425 on the SAT, which was updated this year by the College Board to a 1600-point scale and a new Evidence-Based Writing and Reading section. Last year’s early action admits averaged a 33 on the ACT and a 2128 on the old 2400-point SAT.”
@-------- thanks for posting the links to the EA reports.
one may also be interested in browsing the parallel thread "http://talk.collegeconfidential.com/college-admissions/1961046-old-sat-scores-compared-to-redesigned-sat-scores-p8.html
[noparse] nope. [/noparse]
[noparse]
nope.
[/noparse]
The University of California has posted on its Web site the 25th and 75th percentile scores for the redesigned EBRW and Math SAT tests for the high school class of 2017, for each of its 9 campuses. Using the Internet Wayback machine, we can also retrieve the 25th and 75th percentile scores for the old Critical Reading, Writing, and Math SAT tests for the high school class of 2016.
This may be an instructive data set, since the UC system receives a large number of applications (over 200,000), which helps to reduce variability, and because the statistics for the 9 campuses are apparently being calculated using uniform methodology, and because there is a range of SAT scores among the campuses.
I will focus on the data for the Math test, since we do not know the correlation between the old Critical Reading and Writing test scores, and it is therefore more difficult to match those scores against the EBRW test scores.
Here are the data, where I am using the SAT concordance table between old and new Math scores:
School, Old Math, Expected New Math (via concordance), Actual New Math
Berkeley, 660-780, 690-790, 640-760
Davis, 590-750, 610-770, 600-750
Irvine, 590-740, 610-760, 600-740
Los Angeles, 640-770, 660-785, 640-770
Merced, 470-620, 510-640, 510-620
Riverside, 510-670, 540-700, 540-670
San Diego, 630-770, 650-785, 630-760
Santa Barbara, 600-760, 620-780, 600-750
Santa Cruz, 560-700, 580-730, 590-710
For scores in the 500-590 range, the actual and expected scores from the concordance table match very closely. For scores in the 600-790 range, the actual scores are 10 to 20 points lower than those expected from the concordance table. And for scores at or above 700, the actual scores are 15 to 30 points lower than the expected scores.
In the above message, there is a typo. It should read “For scores in the 600-690 range, …”
Prompted by another post I checked prepscholar’s conversion table and it looks like a 1430 SAT is now a 32 ACT while it used to be a 31 same as the official CB table. It looks to me that people are playing around although next year’s results might again change things.
It appears that the Prepscholar Web site changed its concordance tables some time between August 2 and now. There are significant differences between that conversion table and the 2016 CollegeBoard conversion table. For example, a redesigned SAT score of 1440 is ACT 31 using CollegeBoard but 33 using Prepscholar. Likewise, SAT scores of 1400 and 1410 are ACT 30 using CollegeBoard but 32 using Prepscholar.
The strange thing is that the Prepscholar Web site says “The most accurate concordance tables come directly from the makers of the test – College Board. These are the most up to date tables officially released.” The Web page also says that the table is “via College Board” and that “this conversion chart comes from the College Board only.” However, as shown above, the concordance table currently on the Web site is significantly different from both the 2016 table from CollegeBoard and from the archived copy of the Prepscholar Web site from August 2.
Why would Prepscholar be a source that people use? Makes sense to just go to College Board. While looking for an SAT target score for S19, we decided to go with 1520 because CB says that’s the equivalent to a 34 on the ACT.
It’s not that we want to use Prepscholar. Rather, many of us have a suspicion that the CollegeBoard concordance tables from May 2016 are not accurate, especially at the high range of SAT scores. See, for example, the University of California evidence I presented above, as well as other anecdotal reports coming in from various admissions committees. And now, here is a Web site that has mysteriously changed its concordance tables to support our suspicions, while professing to get its data from the CollegeBoard.
It looks like PrepScholar’s ACT-SAT score converter uses CB’s 2009 document https://research.collegeboard.org/sites/default/files/publications/2012/7/researchnote-2009-40-act-sat-concordance-tables.pdf . Alternatively, I might wonder if CB quietly published new ones, but I’m not finding anything. If these were PrepScholar’s own estimates, I doubt they’d include the statement that the numbers come directly from the College Board. My guess is that someone at PrepScholar made a mistake.
If there are other converters or tables out there that purport to be recent, I’d be interested in looking at them, but I haven’t seen any.
Perhaps this will be moot when the new Common Data Sets are published - wishful thinking?
I believe the SAT data that Williams reported showed that Avg SAT scores for accepted students were slightly lower this year than last, which may mean the original concordance tables by College board were off.