Old SAT scores compared to redesigned SAT scores.

@am9799 Yes, you’d think it would be apparent, but maybe it’s not so easy to see it as it’s happening & in the context of ‘holistic’ admissions. Upthread, I stated my hope that the schools would be smart about this, but it gives me pause that they’re willing to use such absolute language in the common data sets. As you say, we’ll see.

The potential inaccuracy of the concordance tables adds more uncertainty for the student trying to create a college app list. If the tables are wrong, students may under-shoot.

I’m wondering how Naviance handles new vs old scores.

“Do convert New SAT scores (2016) to Old SAT scores using the College Board’s concordance tools and tables”

This seems to be a requirement of the Common Data Set, which makes me wonder whether most schools will really publish scores for New vs Old for the current admissions season. While a few schools may have done so for the early part of the season, that occurred in articles and press releases, yes? I’m still concerned about how much data on admissions with the New SATs will be available this fall - I feel bad for class of 2018.

Editing to add, wait a minute. CDS will be reporting old sat scores. Students with new sat scores are then supposed to use the tables to compare their scores to the CDS? Shouldn’t the CDS have done it the other way around, converting old to new for students going forward?

Naviance added point to old sat as per concordance table. Yes I understand the under shooting and perhaps the intimidation factor. Going back to Northeastern they put at their website an estimated 50% 1440-1540 (from 2010-2200). Got to see if they meet that as it is a school one of my kids is interested.

^ “Going back to Northeastern they put at their website an estimated 50% 1440-1540 (from 2010-2200). Got to see if they meet that as it is a school one of my kids is interested.”

Was that change from Concordance only or from higher stat kids applying to Northeastern each year. Probably a bit of both?

Adcoms aren’t “accepting” the tables in the sense of checking them, case by case. They reviewed the available info and now can compare score strengths within their own applicant pools and the high school or individual context. They know what they felt was a good old score and are learning what’s a good new score, in their pools.

If the new test “graded” you higher, you can’t assume it means you stand higher. The numbers will likely shift upward. Williams and a few others are reporting a jump in the number of Early apps from “high-achieving, low-income students,” which, for now, may explain some of that 7 point drop in R and M. It’s just waay too soon to know how this will shake out.

Adcoms aren’t just gong to say, wow, he got a 7xx, check!, when they know the new scores run higher based on the new formulas.

If someone didn’t take the old SAT, it doesn’t change the fact that the new scores run higher.

Too soon to parse a few new reports.

@lookingforward “If someone didn’t take the old SAT, it doesn’t change the fact that the new scores run higher.”

Is your only source for this “fact” the college board concordance table?

@londondad
Right now from concordance only

@LadyMeowMeow no. I work for a school expecting to see higher new scores than old, based on the new formulas/considerations and what they specifically see in their pools, many different sorts of factors. What they say is: we don’t know the shake out yet and that different top colleges may see different trends now, based on just who’s applying. All those kids who have a new 720, eg, may be applying, forgetting that today’s new 720 is not yesterday’s.

CC tends to like absolutes. A 740 is better than a 710, which is better than a 690. But you’re applying in a new context where so many kids benefitted from the new scoring bumps.

And it’s not just about the totals, but the individual scores. Frankly, if I had a junior, I’d be thinking hard about trying the ACT.

@lookingforward I’m not following your logic as you are assuming that the concordance chart is accurate. Many, including myself, are saying that it doesn’t appear to be the case.

londondad, my daughter only took the new SAT and she was surprised to get into U-Mass Honors.

Here’s my theory. First, to state the obvious, a student who got 780 can not improve by 40 points. So the scale is adjusted for lower increases at the top end. I don’t believe it was adjusted enough, and here’s why. And in case anyone is new to this, the old SAT had 5 possible answers and the new SAT had 4 possible answers. In theory, taking the guess work out of it.

Take student A, Will call him Dad, and he got a 500M on the new SAT. He has a lot of questions he isn’t sure he knows the answer to, so he can make an educated guess on 6-10 questions, and gets more than 25% of those right, because he could rule out 1 to 2 answers. He guesses on 8-10 that he has no clue on and gets 25% right on average.

Take student B, Will call her daughter. She gets 750M on the new SAT. She only has 4 questions she doesn’t know the answer to. She is not able to rule any answers out because these questions are so hard she doesn’t have a clue.
She can guess on these and may get 1 out 4 right, if she even has any time left.

Dad has a better chance to get on additional 6 to 8 questions right by not being penalized for guessing on this years test, but daughter will only get 1 more question right since this years test only has 4 possible answers compared to last years test of 5 possible answers. Guessing actually helps Dad more than it helps Daughter, even more than they thought. College board might have thought empty answers would help them equally, 25% guessed right. Dad had a better percentage at guessing because he could rule out some answers, or just even looking at the answers made one or two look more realistic. Daughter didn’t even have time left to guess, but was pretty happy with how the test went so she felt no need to quickly just bubble the last 4.

So along with a different pool of applicants, less test prep professionals for the new test and my Dad Daughter theory. I believe the concordance tables are off, mostly at the high end.

I assume colleges will take a better look at the percentages.

The 2016-17 Common Data Set will report the old SAT (and convert any new to old scores for the purpose of reporting) as that cohort of students primarily took the old SAT (Class of 2020). The 2017-18 CDS will flip this and report the new SAT (and convert the old to new) because that cohort of students will have primarily taken the new SAT. The schools have to convert (using the SAT conversion charts) for purposes of reporting on the common data set because the CDS has decided to only account for one SAT score in the report. This does not necessarily reflect the view of the admissions committee even within the same institution. Admissions committees will tend to look at the old and new SAT as separate tests, but the common data set will not.

The concordance is just one front-line look-see, an imperfect tool, throwing out some statistics. What matters is how the new scores shift the competition. If more kids can get a 720 (just to pick a reasonable number, not the extreme top end,) then it can’t carry the weight it used to, as a single measure. In old thinking, a 720 may have been great. In the new, it may be easier to achieve.

Try not to be absolute about it.

In a recent enough thread, about what MIT considers good enough, one of their CC members said anything with a 7 in front means you’re “capable” of doing MIT work. As the scores trend higher, because of how the new test is built and scored, I think the true benchmark will increase.

It’s like grade inflation.

Thanks everybody. The thing that I really don’t get is that the old SAT was 1/3rd M, 1/3rd CR and 1/3rd W whereas the new one is 50:50. How is that issue reflected in the concordance tables?

@lookingforward “I work for a school expecting to see higher new scores than old, based on the new formulas/considerations and what they specifically see in their pools, many different sorts of factors.”

This is exactly the scenario that worries parents of kids who only took the new SAT and are now awaiting admissions results: schools “expecting to see” something based on nothing other than a concordance and a few nebulous, unidentifiable ‘considerations’ and ‘sorts of factors.’

A few years later, the data gets ‘corrected’ and hey, turns out @Akqj10 was right, the concordance was wrong, especially at the top end. Ho hum. Wonder what happened to all those top-end new SAT takers?

I’m half kidding, again, but I hope the point is clear.

@lookingforward I’m not being absolute about the new vs old SAT. Just the opposite. I’m supposing that it’s possible that a 720 could be easier to obtain on the new test, but that a 770 could be harder. That’s why I don’t think you can just wave at it and say ‘grade inflation.’

I have real concerns with the concordance table btw New SAT and Old SAT, esp. the top end. I have multiple students got 2150-2250 in the Old SAT AND COME UP short in the New SAT. These kids get roughly 30-50 lower than the concordance table suggest. For example, I got a kid with 720 in all three sections in Old SAT and scored 730M, 720R in the new one (obviously, I told him not to submit the new test data vs the old since the table suggests he should get1500 overall in the new). This has happened more than just once… in fact, multiple times with the kids who took the old and the new. Amazingly, these kids are prepped for the new test and still many have lower scores than the corresponding number from the table. The table should come with a big asterisk for people using that for admission decisions.

Yes, LMM, worrisome, all around. One problem with the tables is they aren’t breaking it out math vs CR, while adcoms can look for the right strengths per the possible major. The total isn’t as important as the components (same with ACT composite.)

But if it’s “easier” to get that 720 than in recent years, it dilutes it. (I hate to say that, I know scores alone aren’t what pushes a kid to an admit, but here we’re just talking scores.) It doesn’t matter what old number the tables equate it to, as much as how it now stands out. Or not. Somewhat subjective and again, depends on the pool.

Take comfort in this: adcoms know it’s screwy, right now. They are not saying, “Oh, the tables show…” They’re looking at the same combinations of factors as before. What your grades, rigor, standing (if reported,) other achievements, all show about you, your stretch, and what ‘could’ be expected in scores. And how others performed and scored (hs, area, national.) And then saying ok, and looking at the rest of the app/supp.

And I’m talking tippy tops.

Btw, the colleges can separately label new vs old, so they see the diffs and can take the best scores into account. No one I know is saying, well the new scores were lower than the tables show they should be. Each can be viewed in its context as old or new, if both are reported. And if the ACT is included, they can go for the strongest set, among all 3 tests.

Maybe I’m just not understanding, but you certainly don’t seem to get a bump at the high end from the new SAT. The comparison engine on the College Board site suggests that if you score 1560 (790V+ 770M) on the new SAT, that’s equivalent to 2300 on the old SAT. If you just multiply 1560 by 1.5, though (to create the equivalent of three sections), you’d get 2340, and if you took the view that because there used to be two “verbal” sections, you should multiply 790 x 2 and add it to 770, you’d get 2350. Am I missing something?

Nope. Many of us think those are wrong. I think they will be adjusted at some point.

One thing that I noticed which seems to be a difference between the way the new and old SAT are scored have to do with the differentiation in the top end of the scores for each section. For example on the old test, 1 or 2 wrong in the math section could drop your score from 50 to 70 points. On the November 2014 test, one wrong on math got you a 750. That’s why on the old test, it was actually kind of rare to see a student score between 2350-2400. I think it might have been more common to score a 2400 than say a 2390.

But on the new test, for the most part it looks like each question is worth 10 points even at the top end. So now it is possible to score a 780 or a 790 on the new test which was actually impossible to do on some administrations of the old test. It appears now that there is a fairly even distribution of scores between 1500 to 1600.

I am not sure how this plays out in terms of the concordance tables and how colleges view it though.