It makes sense because the national percentiles on based on a STUDY GROUP composed of a representative sample of all students in the nation, including students who usually would not take the test.
@Frankmeister ā Thanks for calling and sharing that information - very useful to know - so in many ways CB is leading people on by using National percentiles rather than the research study group ("Userā percentiles) or actual test takers. We know the National percentiles are inflated and if 99+ means 99.1 AND now the definition changed to be "AT or below) then the SI tables are quite off (very āoverstatedā) & really donāt provide very meaningful guidance.
āNationally representative percentiles are derived via a research study sample of U.S. students in 11th grade, weighted to represent all 11th grade U.S. students, regardless of whether they typically take the PSAT/NMSQT and PSAT 10. For example, a studentās score in the 75th percentile means that 75% of the nationally representative group of 11th grade U.S. students would have had scores at or below that studentās score.ā
I see that the āUserā percentiles also are ānationalā but if CB told you they did not base the SI tables on National āUserā percentiles, then we know what is being reported is really not accurate.and honestly it does raise EVEN MORE doubts about CB and their motives.
Ugh!!! They have real data and refuse to report it & so many students & families are trying to make sense of it ā it really feels like mostly a ploy to get more students to take the SAT.
99+ (SI 214) canāt be 99.1% if the mean and SD are correct for p.11 of Understanding Scores.
@AZcat1 please explain post 2742 since Iām not a stat person. Thanks.
@AZcat1 That mean and SD are meaningless. How could the mean be 468 when the max value is 228?
@NathanBN ā They recently changed the understanding your score report to say SI Mean of 148 and SD of 26 on Pg 11:
https://collegereadiness.collegeboard.org/pdf/2015-psat-nmsqt-understanding-scores.pdf
@nathanbn the mean of 468 refers to an SAT score (out of 760) and the 148 refers to the SI (out of 224)
99+ 99+ 99+ percentile meaning 99.1, 99.2,99.3 percentile seems likely, I think we would assume similar progression where 99,99,99 percentile was used. This is giving more granularity compared to scores getting clubbed into the next 0.5 interval.
For SLparent, here is a short tutorial on how to calculate percentile from mean and SD on a normal distribution.
http://www.medfriendly.com/standardscoretopercentileconversion.html
If the mean is 148, and SD is 26 for the SI, then the 99% would start at 205.2 (2.2 x 26 + 148). 99.53% would be 215.6 (2.6 x 26 +148).
Past PSAT scores do not fit a normal distribution, but are only off by 3-4 points moving toward the high end.
So, the supervisorās statement that their definition of 99+ (SI 214) is 99.1% is impossible with a normal distribution. 99.1% should be near 205. Frankly, I would expect the national numbers to be more like a normal distribution, than the actual PSAT test data.
Luke: This new race of people called the Masters claims to have a very large pool of knowledge that will solve the greatest mystery of our time.
Obi-Wan: Luke, use the force and solve the mystery yourself.
Luke: But this knowledge will allow us to destroy the dreaded PSAT, the great slayer of manās hopes and dreams.
Obi-Wan: Luke, use the force.
Luke: But they, ā¦
Obi-Wan: Luke, there is nothing there, move on.
@AZcat1 Thanks!
@AZcat1 Or anyone else! Question: At the risk of sounding ignorant - does this then mean that of the approx 1,700,000 juniors that took the test, only about 7,990 scored at or above SI 215.6?
@CA1543 So CB took the time to update statistical info on page 11, but couldnāt be bothered to correct the scoring error on pg 13?
I know thereās an emoticon around here somewhere to illustrate my disdainā¦
@ HalfMoon 22, afraid not. CB used the National numbers to create p.11 of Understanding Scores. And, the actual PSAT test data will not follow a normal distribution. If 215.6 is 99+ for the National numbers, I would expect 217 for the National-User numbers, and perhaps 220 for the 2017 PSAT. So, 7,990 at or above 220.
anyone know how to hyperlink the names at the beginning of a message?
@SLParent #2642. Thank you, I had read your post (# 2247) with great interest as my son is in the exact same position ā done with testing unless he needs to test for NM. We are lucky to have a good counselor (small school). I didnāt call her to ask her about it as I am taking all the info that comes out of CB as being like going to DMV. You get a different answer depending on who you get on the phone. Maybe not the case but thatās how Iām treating all of their info for now. GC called me to talk about my sonās college prospects and to make sure I knew the test dates for my sonās AP exams and subject matter tests. Thatās when I decided to throw in the question to get her thoughts. Wonderful GC but I make sure to keep myself informed so I can make sure my childās future doesnāt fall through any cracks. Thatās why we are all here reading along. There is a wealth of information on this site for which I am very appreciative. I will still be waiting for the official requirements posted by NM because I think this year everything and anything could change. Itās unbelievable what is happening to this junior class. Off topic but for those sophmore and younger students, college counselors told us there is a ācoalition appā coming (80 schools that will leave the core app) and they were planning on having this junior class use it fall of 2016. College counselors on the advisory board said enough has been done to this class of 2017 so donāt do it. redesigned PSAT, SAT and ACT essay prompts. Enough already.
@AZcat1 - not sure this is what you mean - when trying to tag someone in a comment, a box with username options pops up. Instead of typing out the entire username, click on the one you want. If this isnāt what you mean, then apologies! Thanks for the above info.
@Plotinus - National percentiles make sense by themselves - but itās contradictory to say that the SI chart is a national percentile AND itās based on the people that took the last test (not a study group).
Maybe we could reconcile the SI% table to the concordance table if we believe that the SI% table is based on the national sample and looks at twice as many students as those who actually took the test. For example, for MA in the past few years the cutoff score was the last 99% before the 99+ began. If 99+ is 99.53% as the normal distribution suggests, then you have to be in the top 0.5% to make it in MA. Or in the top 0.25% if you use the National index since they now count twice as many students in the pool. Using the mean and SD provided with the SI% table, we can calculate that 221 is 99.74%. So the cutoff for MA would be 221. The concordance tables concord 220 in 2015 into 220-225 in 2014. The cutoff in MA in 2014 was 223. So 220 would probably just miss it. I know my musings didnāt receive a ton of support on this board in the past, but evidently it didnāt deter me from continuing to try.
@3boystogo ā Yes - basically - I had downloaded the previous version of Understanding Your Scores - in late January - someone else noticed this change before I did. It is pretyy shabby that there is not explanation on pg. 10 or 11 about the source of the SI percentiles - and it is quite frustrating if they used the National rather than User percentiles - there is no justification I can think of for that. Of course they had the actual data and could have used it but suppose they & NM folks did not want to let people figure out too much. :-S