If a school has a tradition of producing relatively fixed number (letâs just say 30) of SF from year to year, and has a relatively large contribution to stateâs SF candidates (letâs just say >= 5%, so in OH 30 out of 600 odd SF) , we can take the bottom SI score of the 30th student and say with reasonable confidence that the cutoff for that state (OH for my kid) will be >= to the 30th student in that high performing school.
Too bad I donât have the data to support this hypothesis. Maybe someone of you well connected student/parent can figure this out for your own state. I donât expect this hypothesis to hold if there isnât an established tradition of putting out relatively constant number of SF and if the school isnât a major contributor to the stateâs total.
@thshadow. I guess Iâm not following you. If this year we have 30 students in the 99% and in the last several years we had to go down into the 98% to get to 8 or 9 students, it seems to me that 99% is not really 99%.
I would not dare compare myself to @thshadow, @Speedy2019 and a few others. We look at CB data table at different angles. But seem like we have something in common - the estimate cut off for 2017 NMF is not as high as Testmasterâs
@thshadow: Iâm probably conflating State and National positions. Last year top 1.15% in CA made sf; but that certainly isnât the same as top 1% on the CB doc, so sitting at 99.6% nationally is probably the exact bubble spot for CAâŠmy problem is convincing myself that it is actually where she is sitting. Even the slightest inflation on the SI% table will likely be a killer for her.
In any event, I am loving following these threadsâŠsuch a great cryptography problem that has made the three weeks since score release fly byâŠonly 30 weeks to go
I think people need to follow up with the counselors as obviously hearing that âeveryone is in the 99th percentileâ is a useless piece of information. What we need to be asking is how many kids over a certain SI. Or what is the lowest SI of the top 10 (20, 30, whatever the number of NMâs you usually have). Only then will you get any relevant info.
I guess GC have limited info and data, your school may have two people with 210âs while other schools and districts may have 100âs in 220âs. Unless you know data for whole state, itâs a shot in the dark.
@hopefulmomof6 - I agree that percentiles seem inflated. (It could just mean that there were more kids taking the test, which would imply that the percentiles may be accurate, but youâd have to have a much higher percentile to get the fixed number of kids who are NMSF.) One thing to point out is that when your GC says â30 are in 99%â - itâs quite possible heâs talking about the national percentile, which isnât really comparable to anything. A 1370 is a 99 national percentile but a (mid) 98 user percentile. Obviously 30 is a big jump. So the tables might show that 1370 is 98.5%. Maybe last year your cutoff was 98.7%, and you had 9 NMSF? Given that you have 30 students with scores >= 1370, maybe that 1370 should really be say a 98.0% or 97.8%. Which would make the 98.5% percentile estimate âoffâ, but I would argue that itâs not way offâŠ
Which would lead me to believe that your daughterâs - did you say 214? - is good enough for your state. My guess is that sheâs probably in the top 5 or 10 for your school, which - if you ignore all the percentile stuff - seems to fit with the estimate of NMSF from your school.
@thshadow So today you have to put a yearâs pay on if the 217 is good enough or not. Which do you put your money on? I bet nay, only because Iâm a natural pessimist.
@thshadow wrote âIt could just mean that there were more kids taking the testâŠâ
1,724,416 juniors took the test this year. 1,595,486 juniors took it last year. That wonât change the percentiles by much.
Alternatively, maybe the SI table is referring to the percentiles out of the approx 3 million juniors in the US. Which, as Iâve stated before, would approximately halve the distance from 100%. Which would mean that 99.5 this year is the same as 99 last year, etc. which again would make the table not smell so bad.
One thing I wonder about is if GCs are getting real percentiles or not - is it from actual test takers, âNationalâ data or âUserâ data?
The National data info in âUnderstanding your Scoresâ has a long range of scores in 99%ile - in the total score table - it goes down to 1390 (User) and 1370 (National) and for the for 99+%ile - it is 1440(User) & 1430 (National)- think only the User percentiles could possibly be relevant & this is still not the actual test-takers. Maybe someone can tweet to the CB and ask if the educator reports are based on real test-takers (we know the studentâs is real & the schoolâs but what about the pool each is being compared to at the state and national level?) - I sort of recall that up-thread there was some indication that it was just based on a research sample - maybe the User data. https://collegereadiness.collegeboard.org/pdf/2015-psat-nmsqt-understanding-scores.pdf If so, then not really saying anything more than a studentâs score report but can perhaps get an idea of the state mean. If we can confirm it is actual state test-taker data & can get studentâs percentile in the state then that is useful info.
I agree with @thshadow. At any rate, just because the high school counselor said they had 30 students in the 99th% doesnât necessarily mean the numbers are wrong because each state has a different cutoff for the exact percentage within the 99% based on the SI and the STATEâs cutoff. The school could have 30 in the 99% but only 9-10 with the SI necessary for NM consideration. Specifics are very important. Iâm sure the actual numbers are correct. There may be more who scored at the bottom 99% but they may be commended or lower cut-off states. We should all take note that 95+ percentiles are an honor and 98-99% are pretty amazing. This test is not easy for MOST of the students who are required to take it, especially where the students are not prepped or in areas where the ACT is king.
Also, I just want to agree with everyone that any PSAT SI or TS in this thread is going to be in the highest 97+ percentiles. Who would bother with research if it wasnât at least partially a consideration for you/your child? And who would post (OR tell their friends- honestly) a score that he/she was not proud of displaying. Even high scorers may not be completely honest if they are surrounded by geniuses.
My son scored a 1410/209 SI in Arkansas (and a 33 on his ACT- for reference). I think he is on the bubble- I am hopeful in comparing this SI to past yearsâs cutoffs. He has several high achieving friends and his score is 100 points higher than theirs- but he is math/ reading strong. He attends a small catholic school where he will graduate with 40 in his class. We have only had one other National Merit Finalist and she graduated with my other son in 2014. She attends Dartmouth where you receive little or no aid from National Merit recognition, just as most Ivys tend to provide need-based or none at all because most applicants score very high on ACT and SAT tests (of course). Just a note to those asking about scholarships: Most NM full rides are State schools, to my knowledge. At present, he is not looking at an Ivy- just an amazing engineering school we can afford. I certainly donât want to come off sounding snobbish
@DoyleB, the SI % table says it is only 11th graders on page 11. So out of 1.6-1.7 million.
May I request of people who post, to use Selection Index % if possible. Saying âmy school has xx students in 99%â means nothing. When talking to principals or GCs, ask for Selection Index percentages. This is the only thing that matters.
Ideally, this kind of information is the most helpful: GC says there are 10 kids in the SI 99% range, last year we had 5 or 20 SFs.
@Speedy2019 Yes there were 1.7 million juniors who took the test. There are approx 3 million juniors in the US - some took the test, and some did not (the nebulous ânational representative groupâ). Does the SI table refer to percentiles from the group who took the test, or the ânational representative groupâ? It makes a big difference in what the percentiles mean, and how to compare them to last year.
Remember, the bullet that describes the population that the SI percentiles are referring to is missing from this yearâs document. Last year it read âPercentiles are based on the Selection Index earned by college-bound juniors who took the PSAT/NMSQT in the previous year.â This year it is nowhere to be found.
Thanks @thshadow for all your work & analysis!!! I think we are in agreement - saw your post at 1490 about what the tables are showing.
@Speedy2019 Not sure the GCs are getting reports showing percentiles for SIâs. Ours report based on total scores.
I just noticed this on Prep Scholar - Dated Sept 2015 but must have been revised right bc they had posted proposed cut offs that were just mins 12 points? Claims to be based on the concordance tables. Matches Test masters I think - http://blog.prepscholar.com/does-your-psat-score-qualify-for-national-merit
âThe average of all the 2016 PSAT NMSQT qualifying scores is a 214. This could look like a 36 in two sections and a 35 in the other or a higher score in one section and lower in another. Sorry to those students in Washington, DC and New Jersey: youâll have to aim significantly higher and achieve the PSAT merit cutoff of at least a 222 on the Selection Index.
Again, these figures are estimates, and thereâs some fluctuation from year to year. To account for this, Iâd recommend aiming for at least 2 to 5 points higher than your stateâs cutoff score.â This was their recent update we all saw: http://blog.prepscholar.com/national-merit-scholarship-cutoff-2015-2016
Hey CCâers â Do we have a chart showing what the historic state cutoffs are in terms of the percentiles of selection indices??
