@CA1543 the College Board website was having trouble yesterday too. A bunch of things were unavailable throughout the day.
@dallaspiano there is the possibility that the new percentile table for SI - from page 11 of the Understanding Scores report - underestimates performance by a few points. It was based on a test sample, not actual data. We already know that actual scores DID outperform the reference group at the mean (a few people were discussing this yesterday on the threads) so it’s reasonable to expect that SI did as well. How the upper tail is affected is anyone’s guess - but at least the concordance tables have a way of referencing back to previous data, whereas right now it’s really difficult to compare current performance to historical norms just by looking at the SI percentages.
@LivinProof wrote:
"OK, wading in here…DD in CA scored a 1460/217 which CB says is 99+% (natl presumably). Assuming the reported % are not completely screwy then I can’t see the CA cutoff being in the 220’s.
Looking at the last two NMSC annual reports would indicate she is safely in:
2014 Nationally - 16,227 finalists out of 1,476,770 entrants = 1.10%
2014 California- 2,027 finalists out of 176,879 entrants = 1.15%
Let me have it…what am I missing?"
Setting aside the potential for bogus SI percentages…
Her national percentile rank is irrelevant, so the fact that she’s in the top 1.1% nationally makes no difference. NMSF are determined on a state by state basis- so you’re correct that what is important is to be in the top 1.15% of CA to qualify. Sadly, you don’t have a CA-only percentile table to look her score up in, so you don’t know her statewide ranking. Thus you don’t know exactly where she stands. She’s probably fine, but you can’t tell for sure.
Really extreme example: Let’s say you live in a tiny state with 50 NMSF. You got your score report back, and you got a 225 - way way up into the 99+% nationally. Hooray! NMSF is in the bag! But unfortunately there’s an exclusive private school in your state full of the children of Nobel prize winners, and they score high. Really high. 10 got 228, 15 got 227, and 25 got a 226. You’re sitting there with your 225. That’s a really good score. But it’s not in the top 1.15% of your state, so you’re out of luck - your state’s cutoff was 226, and you missed it.
Kind of like what happens in DC every year.
Unable to calm my nerve - not a good sleep… Bad, my SAT’s coming this Saturday
But this morning. I woke up and I see the lights
- CB does not perform sampling surveys like our US presidential voting. If they do, it’s still very close with real data
- The CB report was compiled and released to media more than two months after the second test (10/28/2015)
- CB has many scanners and many computers to check and recheck and recheck data for two months
- CB has data on test takers
- CB has information how many test takers in each score slot (ex. 1520 slot, 1500 slot)
- From score lots, CB uses some super unidentified Excel program to spit out TSs, SI, SD, mean, %tile and craps like concordance tables (many issues, many conflicting stories)
- Everything about NMF will go just like @suzyQ7 described (this CC thread concerns the most)
Summary, if you believe CB not having real data and CB picks up data from sampling random surveys, then many fairy tales will come out and they will interpret more tales and tales
@dallaspiano Sometimes I wonder if we are all “stuck in the weeds” - trying to figure out correlations and concordances, etc.
The underlying sample has to support the real data. Isn’t that a premise of basic scientific research? Otherwise all of these scores are not valid.
Your points have much merit based on years of past behavior for both the old old test (prior to 2005) and the old test (2005-2014). Every year, students would look at the percentiles to determine how they performed. Cutoff scores were a little different from year to year, but almost nothing in the 99+. And to my knowledge, percentile scores were always based off of test takers.
Since the test only goes up to 1520, what is a score of 1390 (the lowest 99th percentile) out of 1600?
I agree that a 1520 score is rare.
Once the state summary reports are released we WILL actually know how many students scored in each score slot on a state-wide basis. Hoping those interval slots are something like 10 points or 20 points. 40 would be too wide and would hide a lot of information at the top.
The way CB says something like “representative” like a story about an opened heart surgery patient complimented God and his North Dallas heart surgeon. The surgeon replied " well, you and I are the lucky ones",but deep inside the surgeon knew that it’s not luck, he performed surgeries for more than 20 years and having success rate of 99.85%.
CB says the same thing for very minor reasons (like students forget fill out some things, machine not working right, too hot or too cold the classroom for test takes — very rare for it happens)
Imagine how huge data for US presidential voting, and they can spit out the information of one single voter in US within days after the voting day. Now we have only 1.5 mil test takers, does it take that long to have the results or they still have to use some random sampling surveys from somewhere to put out the results. Likely, it’s not
@DoyleB - your message at #4387 and that of @suzyQ7 now provides clarity that’s so amiss in the posts within this thread, all must read this. In fact, all this talk of concordance only helps you predict whether your child is going to be within the commended lists (nationally) and does nothing to help you predict if your kid makes it into his/her state’s SF nominations. If you are here for the latter, go right ahead and deal with the concordance conundrum. Else, just wait till September to figure out if your kid made it to SF. Short of CB publishing state-by-state %tiles, there is no way to tell.
@Mamelot Take a look at the old state reports - they don’t tell us ANYTHING at the top level, since they don’t break down the top scores of 99%'iles.
I think I meant to say “former” and not “latter”
The state summary reports won’t be that helpful. I looked at a few for our state. Really nothing left to do but wait. I’m stepping away from the PSAT craziness for now and going back to helping my daughter work on her college list. Good luck to everyone with competitive scores waiting to find out.
@suzyQ7 actually the MN state report was very helpful for reading and writing sections last year, as fewer than 1% scored in that top interval (75 - 80). Math was a different matter. But . . .normalized scores are different now and the total and section scores are best delineated in intervals of 10 or 20. I can hope! I think DC will still be tough though.
Of course CB will not want to reveal info. with the state reports so it’s quite possible they use very wide intervals like +40 or something.
@js1029. Sure you’ll have a shot with a 198. I mentioned in a previous post it is all dependent on the number of test takers and the lowest states that cutoff for national merit. People keep saying it is 50000th student. I sincerely doubt that. It’s likely more if anything. If cutoff was 198 and they just stopped at the 50000th score then obviously some kids with 198 would be in and others would be out. How do they determine first priority? Last name? C’mon. I think if 199 covered 49000 students they go to 198 and just cover everyone at that score if it ends up 52,000 so be it. After all they don’t give students a dime for being commended.
Just my opinion but it’s the only logical approach I can think of. I mean what if the bottom of the qualifying students in wyoming or north dakota scores a 197? There might be 60000 scores the same or better than that. In my eyes they’d all be commended. You’d have to confirm with college board/nm but any other approach wouldn’t make sense.
This article here for Houston TX appears to highlight the example @DoyleB was trying to replicate. I wonder if this is available at a national level for every state. See how two schools dominate (almost 52% of all SF’s) in the region. http://www.houstonschoolsurvey.com/national-merit-scholars.html#.Vp-hFfkrKUl
I called my son’s guidance counselor and his score of a 222 from Maine was in the top 4 kids Last year they has 4 semifinalist. you can look on line to see how many kids made it last year and if you ask guidance they may be able to tell if you fell within that number
@DoyleB wrote:
Her national percentile rank is irrelevant, so the fact that she’s in the top 1.1% nationally makes no difference. NMSF are determined on a state by state basis- so you’re correct that what is important is to be in the top 1.15% of CA to qualify. Sadly, you don’t have a CA-only percentile table
Here is where I think natl. % is relevant…it shows she (99+) is in a cohort of 8000 (assuming 1.6mil tests), so if CA had 200k tests then 2000 of that 8000 cohort would have to come from CA for her score to be insufficient…possible but then scores will have to plummet in the other high cutoff states…unlikely.
@OHtoCollege you are correct that people are deriving state cut-offs from national data and that’s a rough estimate at best. There are in fact no hard and fast rules like “top 1/2 percentile on the national scale and you are NM” which applies to every state, especially the very top ones. While the state reports won’t be able to assist much in predicting a cut-off they can at least help everyone understand better what the top looks like. If 2.5% of the kids in NJ scored a 720 or higher in Math, that’s actually helpful information because someone with a 740 might be able to roughly guess that he/she scored in the top 1.25% or higher. And that’s based on actual state results.
Just to clarify, I’m not looking to determine if my 217 DD is nmsf in CA (I suspect she’s a 50-50 bet), I just think the math doesn’t come close to supporting the prediction of a 220+ CA cutoff.
I wonder where I can see the ‘state report’ for boarding school students and internationals? My guess is NOWHERE! I
@LivinProof wrote:“Here is where I think natl. % is relevant…it shows she (99+) is in a cohort of 8000 (assuming 1.6mil tests), so if CA had 200k tests then 2000 of that 8000 cohort would have to come from CA for her score to be insufficient…possible but then scores will have to plummet in the other high cutoff states…unlikely”.
CA has 12.5% of the population of the US. It in the epicenter of the high-tech universe. It has the highest Asian percentage of population of any state in the continental US. It might be unlikely that CA, with 12.5% of the population, could have 25% of the high scores. But it’s certainly possible.