Two sections of Saturday's SAT will not be scored

[wapiti]

There are enough other sections, the College Board said, to provide valid overall SAT scores for those who took the test that day.

The announcement ended an embarrassing day for the College Board -- one in which students traded rumors and complaints, some fearing that the entire test would be declared invalid.

 

On the two sections that will not be scored, some students received 25 minutes of time instead of the standard 20. SAT proctors were given instructions with the correct 20-minute time limit for the sections, which are in mathematics and reading. But the booklets given to students said that they had 25 minutes. At some testing centers, proctors enforced the 20-minute rule, but at others they let the students have 25 minutes. The result was a nonstandardized standardized test for those two sections.

 

The College Board announcement said the following: "After a comprehensive review and statistical analysis, the College Board and [the Educational Testing Service] have determined that the affected sections will not be scored and we will still be able to provide reliable scores for all students who took the SAT on June 6.

 

https://www.insidehighered.com/news/2015/06/09/college-board-wont-score-two-sections-sat

 

Does this sort of situation, where sections are dropped, happen often aside from experimental sections?

 

I wonder what sort of "statistical analysis" took place - perhaps the testers with more time scored significantly higher.  If five extra minutes (25% more time) is statistically significant, I wonder whether the entire test results may overemphasize speed.  Are small increases in speed really predictive of superior college performance and does the CB have research to support that?  Just thinking out loud about the weight of speed vs the weight of answering questions correctly/incorrectly...  Maybe we can't think about relative g-loading the same way we think about it with regard to IQ testing as the SAT isn't really about ability anymore.

 

I wonder what sort of "statistical analysis" took place - perhaps the testers with more time scored significantly higher.  If five extra minutes (25% more time) is statistically significant, I wonder whether the entire test results may overemphasize speed. 

 

I don't think it should surprise anyone that the SAT, especially the math section, emphasizes speed.  It is very common for students, especially non-superstar students, to not finish sections because they run out of time.  A common test-taking strategy is to first find and answer all the "easy" questions in a section, then go back to the "hard' ones.  A good part of test prep isn't teaching the students new material, but teaching them how to answer questions more quickly.

[wapiti]

I have always known that there's some amount of speed element; I guess it's just the relative significance of it that bugs me.  I'm not worried about my kids finishing the math section as much as I'm worried about the reading section except to the extent that the Redesigned SAT adds need for reading speed within the math section.

[snowbeltmom]

I would be livid if one of my kids had taken the SAT last Saturday.  I am finding it hard to believe that the CB can simply not grade two whole sections, yet still have valid results.  Discounting sections puts way too much emphasis on the sections that will be graded, especially in the Reading Section - what one student considers an easy passage, another student may find a difficult passage, yet those same students may have the opposite reactions to a different reading passage that focuses on a different topic.  By having multiple passages, the reading scores balance out.  

 

At the very least, the "non-profit" CB should offer the students who took the SAT last Saturday the option of a retest at no charge. 

 

 

[cave canem]

I would be livid if one of my kids had taken the SAT last Saturday.

Yes, that sums up where I am.

[Caroline]

I wonder if the statistical analysis they did was to see if on older tests they could accurately predict the overall score if they left off those two sections, not if the extra five minutes was making a difference. They haven't had time to score the answer documents from this past Saturday. The data from the old tests is easily accessible for them. Since test prep centers can accurately predict SAT scored based on partial tests, I bet the College Board can, too. I'm not saying I wouldn't be mad if my kid had been taking the test last Saturday. I hope they will offer a retake to those kids. 

[Joules]

I would be livid if one of my kids had taken the SAT last Saturday.

Yep! I'm not happy about it at all. In ds's testing center there was a bit of confusion, but the proctors did the right thing and only gave them 20 minutes, but his sections are cancelled with everyone else's.

 

I am finding it hard to believe that the CB can simply not grade two whole sections, yet still have valid results. Discounting sections puts way too much emphasis on the sections that will be graded, especially in the Reading Section - what one student considers an easy passage, another student may find a difficult passage, yet those same students may have the opposite reactions to a different reading passage that focuses on a different topic. By having multiple passages, the reading scores balance out.

Yep, same with math. Particularly when a child comes out and says that the essay threw him off on the first few sections, but he got his groove back at the end.

 

At the very least, the "non-profit" CB should offer the students who took the SAT last Saturday the option of a retest at no charge.

Yep again. What I would expect from them, like they care, is to go ahead and score these however they have to, and then for students not happy with their scores offer them a free retake in August or September (once schools start and they have facilities.). They could actually just add an additional test for the whole country and make a little money off of it.

 

October is too late for us (and many others), because we are using that date for subject tests and applying early action. I have a friend that did the SAT in May and subject tests in June (just opposite of us). Wishing I'd had a crystal ball and gone that route.

[Joules]

I wonder if the statistical analysis they did was to see if on older tests they could accurately predict the overall score if they left off those two sections, not if the extra five minutes was making a difference. They haven't had time to score the answer documents from this past Saturday. The data from the old tests is easily accessible for them. Since test prep centers can accurately predict SAT scored based on partial tests, I bet the College Board can, too. I'm not saying I wouldn't be mad if my kid had been taking the test last Saturday. I hope they will offer a retake to those kids.

I think it was old tests that they analyzed. There just wasn't enough time to get a representative sample of the Saturday tests in and score and analyze them.

 

What bugs me is, once they decided that they needed to prove these results were reliable, they probably went data snooping and cherry-picked the data that proved them right. Also, even though the distribution may stay the same, individual students may be drastically affected.

 

Here's what I'm wondering, that would particularly affect those going for top scores...If there are fewer problems, how are they going to make just missing one or two equivalent. Missing one might drop someone from an 800 to a 780, now it might drop them to a 740 to keep the shape of the distribution the same. Unless you pull in data from somewhere else (that student's older tests, a problem-to-problem comparison of May's test, students who only got the 20 min?), how mathermatically do you get the fine distinction that every other administration of the test has?

[Caroline]

I think it was old tests that they analyzed. There just wasn't enough time to get a representative sample of the Saturday tests in and score and analyze them.

 

What bugs me is, once they decided that they needed to prove these results were reliable, they probably went data snooping and cherry-picked the data that proved them right. Also, even though the distribution may stay the same, individual students may be drastically affected.

 

Here's what I'm wondering, that would particularly affect those going for top scores...If there are fewer problems, how are they going to make just missing one or two equivalent. Missing one might drop someone from an 800 to a 780, now it might drop them to a 740 to keep the shape of the distribution the same. Unless you pull in data from somewhere else (that student's older tests, a problem-to-problem comparison of May's test, students who only got the 20 min?), how mathermatically do you get the fine distinction that every other administration of the test has?

Honestly? It won't matter what College Board does. Everyone will think their kid's score would have been higher.

[Joules]

Honestly? It won't matter what College Board does. Everyone will think their kid's score would have been higher.

 

That's why I think it would make sense to offer an extra testing session early in the school year (Aug/Sept).  Unfortunately, money speaks louder than sensibility.

 

(It actually seems like it would be in the College Board's best interest, many people do take the ACT in September because October is too late or devoted to subject tests.  Seems like lost potential profit there.)

[plansrme]

My daughter took the SAT on Saturday, and I am LIVID.  This is not the first time we've been screwed over by the fine folks at CB.  She only remembers this error as affecting one section, though, and it was math.  Were there affected sections on both math and English?  That is what the article sounds like.

 

ETA:  Are these two sections in addition to the experimental section?  I've never been clear on whether "ten sections" is actually "ten plus one experimental" or "9 that count plus one experimental."

 

Stupid College Board.

[wendyroo]

ETA: Are these two sections in addition to the experimental section? I've never been clear on whether "ten sections" is actually "ten plus one experimental" or "9 that count plus one experimental."

 

 

The experimental or "unscored" section is one of the ten...always one of the 25 minute sections.

 

Wendy

[Melissa in NC]

Bumping this topic since our scores will be released tomorrow and I am sure this topic will be rehashed.

 

I am not happy about the sections not being scored.

 

<_<

[Joules]

Bumping this topic since our scores will be released tomorrow and I am sure this topic will be rehashed.

 

I am not happy about the sections not being scored.

 

<_<

 

So tired of waiting.  I think one of the concessions should have been getting the scores to us a couple of days early.  I'm really, really hoping they are good (and accepted by colleges with no reservations!)

[plansrme]

So tired of waiting. I think one of the concessions should have been getting the scores to us a couple of days early. I'm really, really hoping they are good (and accepted by colleges with no reservations!)

You should just be grateful that the College Board permitted your child to take its precious test, that they deigned to take your money. Ingrates.

 

(Disgusted with the whole lot of 'em as well.)

[Joules]

You should just be grateful that the College Board permitted your child to take its precious test, that they deigned to take your money. Ingrates.

 

(Disgusted with the whole lot of 'em as well.)

You made me spiit my morning coffee! Seriously, when I've talked to them, I've certainly felt that way.

 

Scores are up. I would like to see his score on the whole thing, but what they have posted is good enough for the cutoffs we were looking at, so we may just be done. (Hopefully they will deign to allow us to take the subject tests in October!)

[plansrme]

You made me spiit my morning coffee! Seriously, when I've talked to them, I've certainly felt that way.

 

Scores are up. I would like to see his score on the whole thing, but what they have posted is good enough for the cutoffs we were looking at, so we may just be done. (Hopefully they will deign to allow us to take the subject tests in October!)

 

Congratulations!

 

My daughter will be taking it again in October.  Her math score was down (but CR and WR were up).  I've explained to her that if she is going to major in math, all of the other math majors will make fun of her for not having an 800.  She knows the material, but I suspect she was a victim of a PSAT-like curve on this one.  One more try in October; she's thrilled.