Macha201210 replies4 threadsRegistered UserNew Member
I don't pretend to understand the nuances of the College Board's curve, but in Dec. 2018 my son got 5 math questions wrong and a 720. For August 2019 he got 6 math questions wrong and a 650!
evergreen51462 replies31 threadsRegistered UserSenior Member
If you purchase the SAS, you can count the number of hard, medium, and easy math questions. The number of hard questions has dropped significantly for all new, not-previously-used forms administered since June 2018. This is an intentional choice by the College Board.
@MommyCoqui I read somewhere that an easier math test helps the lower scorers which are the big majority of the testers. Also controls for not too many high scores if -1 is 770. College Board is a business and I have to assume that an easy math test brings more customers. It really hurts the subsection of the high scorers that are also prone to careless mistakes and I assume that it is a small number of kids and College Board does not really care. I think they benefit having a reputation for an easy test.
My daughter's scores are not available yet but this does not bode well for her as she struggles in math and guessed on many questions. Could the scoring be affected by the lower number of test takers in August? How did people's scores vary for August compared to other tests dates on the other sections of the August SAT?
lbpearso1 replies0 threadsRegistered UserNew Member
My son, a high school senior, studied all summer as this was the last chance to improve his scores before college applications. He cut the number of wrong answers in half (only 6 wrong) and only went up 20 points to a 680. Based on past reults, his score should have been a 710-720.
The math curve for the August 2019 SAT is UNNECCESSARILY HARSH and is screwing over kids who are about to apply to college. We are upset that he did not take the ACT. Let's let the College Board know how we feel and demand they reevaluate this curve:
DadInBoston1 replies0 threadsRegistered UserNew Member
Our son was very disappointed to see he reduced his incorrect math questions from 12 to 6, but his score only went up 20 points to 680. Based on past tests - it should have been 710-720. Everyone should call (866-756-7346) or email ([email protected]) to express frustration and say we will only take the ACT going forward. The only way the College Board will consider a change - which I know is unlikely - is if they think this will affect their $revenue$ going forward.
SJ821814 replies0 threadsRegistered UserNew Member
I get the frustration but a test taker getting less wrong from one test to another may well be due to the fact that the test is easier. With the test getting easier, it is important to focus on avoiding simple mistake. My son took it in August and his first impression was that the math seems to be on the easier side compared to what he had been practicing. It was only his 2nd time taking and he made dump mistake first time around.
My daughter's scores are not available yet but this does not bode well for her as she struggles in math and guessed on many questions. Could the scoring be affected by the lower number of test takers in August? How did people's scores vary for August compared to other tests dates on the other sections of the August SAT?
The "curve" is based on the difficult of the test, not the number of test takers.
If you perform better than 97% of the typical test taking population on an SAT math test, you will get a 760. It doesn’t matter if you got two questions wrong or 12 questions wrong. The statistical process of equating will adjust the raw to scaled score conversion table based on the difficulty of the test.
710 with 4 wrong means about 7% of the test takers had 3 or fewer wrong. If the test had been harder, and the same student got 10 wrong, but only 7% of the test takers had 9 or fewer wrong, the score would still be a 710. They performed exactly the same, relatively, and received the same score.
Yes, silly errors not based on actual ability, or just randomly guessing, will be affected. That’s why CB attempts to make test of equal difficulty, but it’s not that easy to do.
The idea of a “harsh” curve where *everyone* gets a lower score, just isn’t reality.
I don’t see many people complaining when their scaled score goes up with the same raw score. Which would happen by just taking the exact same tests in a different sequence.
evergreen51462 replies31 threadsRegistered UserSenior Member
edited September 8
The problem is that the equating process does not work if test forms are not "as alike as possible" in difficulty, per the College Board technical manual. At some difference in difficulty, there will not be proper standardization. So for example, if you gave a 5th grade math test to a bunch of high school seniors, it wouldn't be proper to attempt to equate between that test and a test with 12th grade material.
In the case of the New SAT, the actual number of hard math problems dropped by almost half between pre-June 2018 test forms and post; these can be counted via the difficulty designations on the SAS, something like going from 28 to 15. From what I understand, the tests pre-June 2018 apparently had around the same number of hard problems as each other, and the tests post-June 2018 have had around the same number of hard problems as each other. Without any public announcement, College Board deliberately reduced the number of hard math problems beginning June 2018, knowing it would make for a severe curve at the top end of the scale. This calls into question how close College Board has come to the line of improper equating (a very serious lack of quality in their work product) and their purposeful intent in reducing the difficulty so dramatically causes one to ponder what outcome they are so obviously trying to engineer.
If any of you bought the SAS, count the number of questions designated as "hard" in the two math sections.
(Footnote, any tests that were administered after June 2018 but were a major reuse from the prior school year would be counted as pre-June 2018 test forms. One example was Aug 2018 and I think there were a couple of others.)
homerdog4967 replies89 threadsRegistered UserSenior Member
so the question is...what is the studying strategy for the high scoring student who needs to make sure that they minimize mistakes? I know some high stats kids who thought the math section was easy but then rushed through and had way too much time to spare. I suppose one thing is to make sure you're taking your time. I'm not so into re-checking problems if done early. Kids don't know which problems to even check and would very rarely change an answer. Are there practice tests out there that reflect this change in the number of hard questions? It would great to practice with those.
evergreen51462 replies31 threadsRegistered UserSenior Member
There is only one officially released (#10 on the College Board site, the test administered Oct 2018). There are others available elsewhere on the internet that have been released QAS.
Replies to: SAT Math Scores for August 2019 Test
The math curve for the August 2019 SAT is UNNECCESSARILY HARSH and is screwing over kids who are about to apply to college. We are upset that he did not take the ACT. Let's let the College Board know how we feel and demand they reevaluate this curve:
[email protected]@collegeboard.org
1-866-756-7789
The "curve" is based on the difficult of the test, not the number of test takers.
710 with 4 wrong means about 7% of the test takers had 3 or fewer wrong. If the test had been harder, and the same student got 10 wrong, but only 7% of the test takers had 9 or fewer wrong, the score would still be a 710. They performed exactly the same, relatively, and received the same score.
Yes, silly errors not based on actual ability, or just randomly guessing, will be affected. That’s why CB attempts to make test of equal difficulty, but it’s not that easy to do.
The idea of a “harsh” curve where *everyone* gets a lower score, just isn’t reality.
I don’t see many people complaining when their scaled score goes up with the same raw score. Which would happen by just taking the exact same tests in a different sequence.
In the case of the New SAT, the actual number of hard math problems dropped by almost half between pre-June 2018 test forms and post; these can be counted via the difficulty designations on the SAS, something like going from 28 to 15. From what I understand, the tests pre-June 2018 apparently had around the same number of hard problems as each other, and the tests post-June 2018 have had around the same number of hard problems as each other. Without any public announcement, College Board deliberately reduced the number of hard math problems beginning June 2018, knowing it would make for a severe curve at the top end of the scale. This calls into question how close College Board has come to the line of improper equating (a very serious lack of quality in their work product) and their purposeful intent in reducing the difficulty so dramatically causes one to ponder what outcome they are so obviously trying to engineer.
If any of you bought the SAS, count the number of questions designated as "hard" in the two math sections.
(Footnote, any tests that were administered after June 2018 but were a major reuse from the prior school year would be counted as pre-June 2018 test forms. One example was Aug 2018 and I think there were a couple of others.)