National Merit Cutoff Predictions Class of 2017

@micgeaux My D got 212 also ( I think you are from TX too?). Where did you get the percentile for the SI 212 because I don’t see it… Only saws % for scores

Please don’t think this is scientific, but for those who want to do the same calculation that I’m doing:

StateCutoffFor2015 = ((2014StateSICutoff - 2014SIMean) / 2014SIStdDev) * 2015StdDev) + 2015SIMean

2014StateSICutoff is the state in question available from http://www.fairtest.org/sites/default/files/NationalMerit2016CutoffRelease.pdf

2014SIMean = 141.9
2014SIStdDev = 30.7
from http://www.bernardsboe.com/UserFiles/Servers/Server_3096886/File/Jill%20Shadis/Ridge%20Counseling/Standardized%20Testing/Understanding%202014%20PSAT-NMSQT%20Scores.pdf

2015SIStdDev = 28 from @billchu2 's curve fitting http://talk.collegeconfidential.com/discussion/comment/19180539/#Comment_19180539
2015SIMean = 146 from page 11 in https://collegereadiness.collegeboard.org/pdf/2015-psat-nmsqt-understanding-scores.pdf

StateCutoffSIFor2015 = (((2014StateSICutoff - 141.9) / 30.7) * 28) + 146

There are many sources of errors in this simple model. As others have said, maybe on a different thread, there can be ± 2 from year to year for a state. Also, we are looking at the right end tail of a distribution, assuming it is normal when it might be skewed.

So, for example, for CA, using my stddev doodle, I get 220 . Using DC & outside US, 225 in 2014, the highest, I get 222. For Illinois, I get 213.

This is just an exercise, probably not that accurate. Please don’t think it makes or breaks your chances.

^^ MN is 212. Testmasters predicted 215. Last year 214.

Would you say 200 is the commended cut off this year? Since it’s the last score that is considered as 97%? And btw do we know how many juniors took the PSAT this year

@Dave_N and @billchu2 can you please remind me (and probably everyone else!) why you needed to do a curve fitting to arrive at the Std. Dev. of 28? Was it because the Mean/SD #'s on page 11 of the CB report is erroneous? Thinking we’d want to use actual mean and SD #'s. The 146 might be fairly close to the actual mean (it’s at the 50%, after all) but the SD might shift a litle (no offense to billchu2 who provided what looks like a very close estimate).

WRT the stddev calculations - I seriously doubt that the curve is normal at the top. (For example I bet there are more 228’s than 227’s.) Also, as I describe below, what happens to scores in the middle seems to be different than what happens to scores at the top.

If you compare the recently released percentile tables (including the 99+), and map it to 2014, you might end up with the following as a way to convert your 2015 score into a corresponding one from 2014:
If your 2015 score is 228, add 12 to get a corresponding 2014 score of 240.
If your 2015 score is 214, add 10 to get a corresponding 2014 score of 224.
If your 2015 score is 205, add 8 to get a corresponding 2014 score of 213.
If your 2015 score is 202, add 4 to get a corresponding 2014 score of 206.
If your 2015 score is 200, add 2 to get a corresponding 2014 score of 202.
If your 2015 score is 198, add 0 to get a corresponding 2014 score of 198.

IOW, the higher your score, the more points you add.

This would imply that the CA cutoff is 213. (Because adding 10 gives you a 223, which was the 2014 cutoff.) Which is obviously at odds with the data you get from the concordance tables. (They imply much higher cutoffs.)

So - I think the percentile tables imply a 213 cutoff for CA, while the concordance tables seem to imply a 219 or 220 cutoff. It remains to be seen which one is more accurate, and/or why they disagree.

@SLParent Page 11 of the Understanding your Scores 2015 report which someone posted a link to lists the score and the corresponding percentage. So 214 is the last 99+ and 205 is the last 99. Based on historic cutoffs, TX has never been in the 99+.

I think you can fall into one of two camps: the concordance camp with sophisticated math models which from what I have gathered, are closely tied to testmasters predictions.

Or, the page 11 camp, which doesn’t concord to PSAT 2014 (but I am just going by all the posts and I don’t understand it). So if you don’t think CB will revise page 11, and this is the official understanding your scores for 2015, than I might be wrong, but NO state required a 99+ to qualify (maybe DC???) So, if you say that all of the states pretty much act the same as they did in prior years, CA, DC, etc. will STILL be the high cut off states, but they still won’t need above a 214 to qualify. (some people have stated that maybe the states will be a little different because this test is more associated with common core, but who knows?) So if your state qualifying score is normally close to the commended score, you will need a 200. If your state is like LA and just under the 99th percentile, you will need AROUND a 204. And TX should be south of 214.

I am not a math wizard, but it seems like the compression took place near the predicted to actual commended. On a sliding scale, commended was expected to be 192. But, it appears that it will be 200.

My earlier predictions for TX were based on scoring two 98th percentiles and one 99th percentile. I reviewed prior years subject scores and in almost every year, you would either qualify or maybe be down one point in TX if two of your subject scores had the top 98th percentile and one subject score had the lowest 99th percentile. Yes, very unscientific, but my method predicted the commended to be 202, which is three 95 percents - one in each subject.

Dave_N’s formula above gives Commended at 200.81… 201.

@Mamelot Page 11 says Mean Score 468 and Standard Dev 94.
Obviously Neither S.I. nor PSAT score has Mean at 468.

Hi, you seem knowledgeable. The lowest 99+ percentile in 2014 was 224 and Maryland’s cutoff was 221. This year, the lowest 99+ percentile was 214. Given this statistic , would my 213 have a shot at NMSF in Maryland.

@payn4ward it’s a cut and paste error from page 9.

@Mamelot OK. Does it say actual S.I. Mean somewhere else?

@payn4ward no - which I think is why Dave_N used the 50% number of 146. I was just confirming that. Too bad CB data checkers goofed up and didn’t notice the error. You’d think they would have double checked THAT one!

Absolutely - based on the percentiles. Based on the concordance tables - probably not. (No one knows why they disagree.)

IOW - you definitely maybe have a shot… :slight_smile:

I think I will have at least May of this year to really understand the cutoff? What we know in may…I know in Sept the school already receives name of all the students that are semi finalist. I am in borderline score, the long wait is not nice! The SAT score or PSAT score comes out fast enough, so what does it take to calculate who are in that certain percentile in each state? Is this not all about math based on data that people already have…what I am missing!

@Eminem99 on the page of the College Board report (page 12/32), just before the SI percentiles, it says that approximately 1.5 million juniors are NMSC participants this year.

Does anyone think that the percentiles in the new SI percentiles chart may be our best bet at predicting cut offs? I don’t understand why people think the cutoffs may be so high. If a 205 is the 99th, wouldn’t that push down even the highest cutoffs to like 216?

I think that the percentiles are usually based on last year’s test takers. The SI percentiles chart, while nice to see where they think 99 vs 99+ falls, is only based on a “representative sample.” I don’t feel I can rely on College Board’s ability to generate a representative sample to the level of precision necessary to guess at the 16,000 cutoff nationally. I’m going to wait for the official commended cutoff and hopefully some percentiles based on a lot of students to think that we can make firm predictions.

@micgeaux thanks for your response. Where did you get TX’s history of percentiles for NMSF? I would love to see that if you can show me. Do you have older children so you got info from previous years? This is all new to me. Also, are these percentiles user or national? User %'s are more stringent right?

@SLparent No, my twins are my oldest. If you google, you can find historical cutoffs. There is a nice table in one on college confidential threads. It is for test years around 2003-2012. You can then add the data for 2013 and 2014. Keep in my that the test year is two off from the graduation year. So, cutoffs for the class of 2016 are really test year 2014. That may seem self-explanatory, but I think the chart that I found on cc was was posted in August 2013 and it had 2013 listed on it. So, I surmised that is was test year 2011 and graduation year 2013.

From what I’ve read on cc, the user percentiles are the ones to use. However, I think people are trying to figure out which students are in the sample and if this is IT (MEANING THE FINAL). In prior years, the percentiles were based off of actual students who took the test in the PRECEDING year. The sample size gradually increased each year, and if I remember correctly the sample size for the 2014 scores were based on over 1,000,000 test takers in 2013. But, this year there is no prior test like the one the students took in 2015. So, I’m guessing that is why the concordance was made. So, ideally, the 2015 scores SHOULD concord with 2014, which should correspond to the percentile chart on page 11. But alas, it does not. I think another poster remarked that if you have ALL of this data in a database, why can’t you just force rank all of the scores and lop off the top 1%, 2%, etc. I’m sure there is an explanation, but I don’t know.

My only hope is that the concordance tables say preliminary while the percentile tables do not. And, if they revamp the percentile tables, do all the students then get new score reports? (I doubt that.) For example, my son’s writing score is 34 at 98th percentile. The next percentile is 33 at, I believe the 96th percentiles. So if you redo percentiles and they go down, do they issue my son another report saying his writing score is now 97th percentile? Or, do they say we revamped the percentiles for the selection index (which currently doesn’t have a percentile on the report), and don’t worry about the section percentiles.

So, I took the cutoff scores for TX in 2003-2014 and compared those to the lowest 99% selection index to see where it fell, but it changed from year to year. I also took the qualifying score and compared it to earning the lowest 99 in one subject and the highest 98 in two subjects. Sometimes, it equaled; sometimes it fell slightly short (one point).