It doesn’t look to me like much thought went into these estimates by Prepscholar. All states with cutoffs of 209-214 last year have just been lumped together and given a predicted cutoff score of 215 this year.
The lumping together is what the concordance tables do. Testmasters fit a linear relationship to the results, so they don’t show the same step function in their cutoffs, but if you look at the testmasters charts you’ll see the same thing.
@VandyAlum93 Maybe that’s because of the score compression?
Can we just ignore these new predictions? They are based off the concordance tables, tables National Merit said specifically not to use to try to estimate cutoffs as they were meant for sophomores planning for next year.
@garyasho2 the thing to remember is that the concordance tables concord to “previous” tests w/o specifying which one. It’s not incorrect to use class of 2016, or class of 2015 PSAT cut-offs as long as they don’t move around that much. It would probably be a good idea to concord to both sets of old cut-offs just in case so you have a good idea of the maximum range of values (which would be preliminary all the same, btw). I’m betting that the tables are also less accurate overall if there is significant movement from one year to another, as you see in some of the higher-scoring states or selection areas. The problem there is that the movement might have continued even under the old test. I’m not sure the concordance tables capture that dynamic aspect - they are probably much better applied to cut-offs that are a bit more static.
Of course, Prepscholar’s revised predictions are looking more like Testmaster’s now because they are both using the same source to predict and that is the PRELIMINARY concordance tables.
Can someone please call National Merit Scholarship corp tomorrow to ask if we should be using the concordance tables as a means to predict national merits? They told me not to use the tables for that purpose but I’d love to see if someone is told differently than me.
They aren’t going to say that it’s ok because their method of choosing NM’s is totally different from what we, and Testmasters, and PrepScholar do. Can you imagine if they say it’s ok, and then the actual cut-off turns out to be very different? Ouch.
@DoyleB, I think the linear approximation that Testmasters used seems much more reasonable. It is hard to believe that there will be no states with cutoffs between 211 and 214 as Prepscholars suggests with their stepwise model. And can someone explain to me how they are pulling those numbers from the concordance tables? How does KY’s cutoff of 210 last year become a predicted 215 for this year?
Florida’s cut-off for last year is wrong, too. Unless they refer to PSAT 2013 or PSAT 2012. Old can mean any year prior to 2015.
On another note, does anyone know how different PSAT 8/9 and PSAT 10 are from PSAT 11? In the past, all PSATs were the same, no matter when a student took it. When they say 4 million students took the PSAT, they must include all different versions. Florida makes all public school 10th graders take the PSAT on the Wednesday administration day. At my daughter’s school only 8 to 12 11th graders even took the PSAT; the school usually has 4 or 5 NMSF.
Wow I haven’t checked this in a while and the first thing I see is a predicted score of 215 in Kentucky. Is there any hope left for me with a 210 in Kentucky or should I just assume I’m screwed?
So despite being told to NOT use concordance tables to estimate cutoffs every prediction uses the concordance…hmmm now what was that definition of insanity again?
@DoyleB “Not really. A 23 point difference is fairly significant. 987 is about the 52nd percentile; 1010 is the 56th. It means the actual test takers outperformed the reference population by a fair amount.”
That doesn’t seem that earthshaking to me and it seems to fit with that bulge along the curve that was so well displayed earlier in the thread. I think that difference is absorbed well before reaching the tail.
@SLparent If you want to know how NMS calculates commended and finalists, read this post:
http://talk.collegeconfidential.com/discussion/comment/19197995/#Comment_19197995
Hi Everyone…How does a 210 look for TN? Haven’t been on this forum for a while. Thank you.
Re the Prep Scholar cut -offs - they must ave used a prior year (not 2015 as they stated). http://www.fairtest.org/sites/default/files/NationalMeritSemifinalist2016-scores.pdf
I don’t think they can be very precise - almost better/safer to give a 3-4 point range. But I doubt we’ll get any more clarity real soon. I imagine many will be surprised that the top score went down 12 points but their state cut off actually went up - and wonder what went on with this test. Of course so much analysis has been done on the CC threads by some great contributors that it now seems a logical result. If anyone gets to see the info the GC’s can access in their reports please share what you can.
How is a 220 for Illinois?
so i guess a 212 in california is out of the question at this point lol
My humble 2 cents opinions
After reading new estimates for cut off NMF, it 's a mixed feeling. Since CB (in business for more 50 years) they know their data, they know what is the lowest 99+%tile (99.54). @suzyQ7 has outlined their process.
There is no way that lowest 99+%tile go up north like new estimates - about 99.85. The % and scores are translated to SI, percentile, SD and etc.
Imagine with number 99.85, mean that many students (at least 7500) have to score 1490 and up to max 1520. Yes I am aware of perfect scorers - very rare.
Based SAT of 2014 and 2015,
-
around 2500 out of 1.65 mil got that far at 99.85 (score 2350 to max 2400). Do you know how hard it is to get 2350 out of 2400 and to be at 99.85%tile
-
around 7500+ out of 1.65 mil got it from lowest 99.54 to max of 99.99
The reason I use SAT (2400) data since it behaves just like PSAT data (240)
Now, all sudden in one year, this 2015 PSAT it balloons up from 2500 to 7500 (at that 99.85%tile and up)
Yes anything can happens, Statistic is logic and a science. It bases on real data in order to predict. There is a little chance for CB (company with more 50 years in business) will make mistakes and there is no way CB can not see astronomical jump up north with Total score and percent in certain scores (translate to SI)
Data from @LivinProof with post#4308, would support
@dallaspiano so in your opinion based off of what you said, how much lower do you think the cutoff a will be relative to what testmasters is predicting
On College Board PSAT pg now:
Temporarily Unavailable
Score Reports are temporarily unavailable. We apologize for the inconvenience.
Try going back to the main home page.
Maybe they are uploading more scores??