College Board Error on June 6, 2015 SAT?

It would appear that some people don’t understand that the College Board is the one with expertise in testing science. Once they indicate what the solution is, it would be up to an outside entity to challenge that claim.

Unfortunately, nearly all of the vitriol here and in other threads has to do with fear and disappointment rather than with any real understanding of how tests like this work.

So will this be applicable to only students in the United States or internationally?

@JustOneDad Some of us are calling for the College Board to provide proof of it’s statements. Have you been presented with any information verifying that scores are still accurate (actually CB still keeps using the word “reliable” not accurate)? I would love to see their breakdowns on other tests that have had a similar complication. It would also be very easy for them to take a selection of older tests that were administered under normal conditions and recalculate their scores leaving two sections out. I’d love to see a side by side comparison of what the scores were before and after. That would go a long way to help subdue the fear and disappointment that people are naturally prone to in a situation like this.

@Aleksandr7 That is such a good point. I took the SAT as a Frenchman, and scored a 1710, if I improve significantly over that, colleges might just think that I benefited from the curve.

In response to JustOneDad, I made that point in long conversations with ETS and CB. I asked for data for the June test for five years. They are supposed to “get back” to me.

@WaltzingM The College Board routinely makes data available to qualified researchers. The results of that work are made available on their website. Picking apart how these results differ in minute detail from other historical results will no doubt be the subject of a lot of discussion among eggheads in the future.

However, because nearly all the test-takers have no idea of whether they would have gotten a 640 or a 650 or a 630 and how it would even affect their admissions if they had, whining about it seems ridiculous.

@JustOneDad
“However, because nearly all the test-takers have no idea of whether they would have gotten a 640 or a 650 or a 630 and how it would even affect their admissions if they had, whining about it seems ridiculous.”

That is the exact reason why the other sections should be scored and the resulting total score given to the test takers- irregardless of whether CB decides to not count these sections in the final test results. Test takers deserve to know the impact that the change in scoring had on their scores.

I am just curious as an adult out of the loop on SAT’s and following this… are all sections equal… meaning in Math for instance… are all sections multiple choice or are some sections open-ended? if they are keeping open-ended sections and getting rid of mult. choice how does that skew results if you can’t guess… or maybe this is a non-issue?

I guess my concern for students would be that this is a regarded as a standardized test… however if a small subset of college students are being scored on a different “Standard” ie. 4 sections and not 6 like all other students… how is it now a valid “Standardized” test… by definition it does not hold up. What will colleges think of that if they can not hold the scores up to the normal standard? just curious…

*** also… I guess for the students … would a free retake in June be something that you would take advantage of or do most of you just want to take the score and try again in the fall?

@JustOneDad The lack of compassion is remarkable. I may not be a qualified researcher, but I am a parent who paid for a service that failed in part due to a simple typo that should have been caught in any one of the multiple stages of editing these tests go through. They say they gather enough data to capture and model missing sections (even for outliers), but it would be nice to have some proof. If they have qualified researchers studying their data and supporting their claims, who are they? I’d like to hear from them. That’s not whining. That’s a demand for decent customer service.

This doesn’t just affect admissions, and small differences in scores can be a big deal for individual families. Our state U bases scholarship awards on these scores. Just plugging numbers into the online scholarship calculator, a 10 point difference in CR+M score can change a scholarship award by $6000 over four years. Right or wrong, fair or unfair, a lot rides on the SAT and ACT and kids deserve a score they can at least trust is accurate. I’m not thrilled to have paid over $50 for something approximately reliable. Statistics may eventually tell us the results will be, in aggregate, the same as what they would have been without the error, but individuals may still be negatively affected.

I don’t think it has been established that the service you “paid for”, did, in fact, fail.

If you have the statistics education and experience, I suggest you begin by reviewing existing work on SAT measurement and validity.

This is absolutely infuriating. We don’t want “reliable” scores, we want OUR scores. I, as well as many other people, did NOT receive the extra 5 minutes. We were explicitly told that the sections affected will be 20 minutes long. Thus, testing conditions did not differ in any way from previous dates, yet are scores are being jeopardized. I felt as if my scores on the latter portion of the exam, especially for CR, were better than in the beginning of the test. What is most disappointing to me is that I felt extremely confident on the math portion of the test but left only one question blank on the second math section. This normally should translate to a 790 or possibly a 780. Now it may translate to below a 750, significantly lower than what I would have gotten. Also, the first CR section was uncharacteristically difficult for me, and that section will now be weighted higher, purely due to CB’s mistake. Another thing I am wondering is how the un-scored experimental section will be factored into this? I personally feel that they should release both the “reliable” score and the score you actually received had this issue not been discovered.

If you feel so strongly about this, I would suggest individually demanding invalidation and a retake.

However, I have to point out that a simple feeling of confidence on your part is extremely subjective.

A single digit change in a lottery ticket can result in millions of dollars of winnings.

@JustOneDad I would argue that the service did in fact fail, and yes, we do pay a fee to be able to take these tests. What happened here is that a data driven entity, the College Board and their associated testing partners, introduced an uncontrolled variable into what is supposed to be a tightly controlled testing environment. The burden of proof is now on the College Board to prove that this uncontrolled variable has NOT contaminated the results. All we are asking for is some correlating data to prove that the scores that will be provided in a few weeks are representative of the results that would have been achieved had this error not happened. Again, the burden of proof is on the College Board to validate the results, since they were the one who failed to provide the testing environment that they promised.

“A single digit change in a lottery ticket can result in millions of dollars of winnings.”

Which is exactly why it’s important to those who choose to play that the system for determining the winner is accurate.

The supporting data is already there. How many of you actually understand the structure and validity of this test in the first place? What the College Board CANNOT do is tutor each person enough to understand it.

It would be quite a plum for some bright young researcher to be able to prove that the CB was deliberately pulling the wool over thousands of test-takers and their families.

“the College Board and their associated testing partners, introduced an uncontrolled variable…”

It’s called a mistake. They happen.

“It’s called a mistake. They happen.”
We understand that it is a mistake and it happens; however, what they do after the mistake is what is infuriating.
I agree with @ekI2498 that they should release both the “reliable” score and the score you actually received had this issue not been discovered. They should also provide a free retake for anyone planning to retake.

I also don’t believe that you should compare this situation with a lottery ticket. Thousands of kids prepare for these tests and should not feel that it is all based on a luck similar to a lottery ticket. If it is, then the system is flawed.

@JustOneDad but how is this standardized test even valid if the standard is being changed for a subset? Doesn’t changing the criteria for scoring make it invalid in concept?

I agree @lonetreegrad if the College Board and others feel it really is not that big of deal then the College Board she release the full original Score and the skewed new score created with the new scoring and if they don’t make a difference than they should have no qualms of being fully transparent.

However, if they do not provide both scores… there is a VERY good reason not to do so. Do not forget this is big business… MONEY always rules decisions. Always.