@Mamelot I know how to check individual scores for concordance but I don’t know a good, systematic procedure to test this chart against the concordance charts. Any ideas?
@liketowrite, good ways. Got to see in different angles.
But if we talk about %tile range. Percentile is percentile (as mentioned by @mantua in post #1894) it does not change with scores up or scores down - at least with CB. I keep tap CB (%tile) for the last 3 years, the lowest of 99+%tile run from 99.52 to 99.54. And I believe CB know their percentiles (to be specific at 99+ range).
It does not hurt to see your way at all, but it conflicts with CB data. But anyway, express your way to find cut off for TX, and CA or any state of your choice.
@dallaspiano I would argue that “percentile is not percentile”. In the Compass education report, they give 2 different definitions of percentile. The SAT previously used one (Definition A) and the ACT used another (Definition B). In this redesigned 2015 PSAT, CB decided to use B.
Definition A: The percentage of students scoring below you.
Definition B: The percentage of students scoring at or below your score.
I moved the percentiles up one percentile point bc of the definition change.
OK, But anyway, express your way to find cut off for TX, and CA or any state of your choice.
l@liketowrite. Thank you
Just go back to my record 2014, and 2015. I see what you mean
2014 definition of “Your PSAT/NMSQT National Percentile”
Your percentile indicates the percentage of 10th grade College Board U.S. test takers who would have had scores below your score
2015 definition of “Your PSAT/NMSQT User Percentile - National”
Your percentile indicates the percentage of a typical group of 11th grade College Board U.S. test takers who would have had scores at or below your score
But please, express your way to find cut off for TX, and CA or any state of your choice.
Can we explore the word “or”?
My basic English tells me that “you can go West or you can go East” does not mean you go East and West at same time
In 2015 definition of “Your PSAT/NMSQT User Percentile - National”
Your percentile indicates the percentage of a typical group of 11th grade College Board U.S. test takers who would have had scores at or below your score
I understand simply
1- There are many students have same scores like myself (TS or SI)
2- There are many students have scores below myself (TS or SI)
If I do not understand it fully, please ignore. Sorry, my bad
@dallaspiano I can’t figure out how to use my chart to predict TX, which last year was a 220 and was 2/3 the way up the 99s because the part of my chart about which I am most confused is how to adjust the 99s and the 99+s.
I think the def is more ‘and’ than ‘or’. If you have not read the article, you should. http://www.■■■■■■■■■■■■■■■/wp-content/uploads/2016/01/Problems-with-the-New-PSAT-Compass-1-26-2016.pdf
@Mamelot I went back and saw your post. I think when I read it yesterday, I thought “up a notch” was a single SI but I guess you meant a single percentile. So is this why everyone started giving lower predictions and then predictions similar to last year? I could not figure it out but now I understand why. But, do you have any idea how to adjust the chart in a reasonable way that determines where the split is bw 99s and 99+s?
I can not edit my post, so again
Can we explore the word “or”?
My basic English tells me that “you can go West or you can go East” does not mean you go East and West at same time
In 2015 definition of “Your PSAT/NMSQT User Percentile - National”
Your percentile indicates the percentage of a typical group of 11th grade College Board U.S. test takers who would have had scores at or below your score
I understand simply
1- There are many students have same scores like my score (TS or SI)
2- There are many students have scores below my score (TS or SI)
If I do not understand it fully, please ignore. Sorry, my bad
@dallaspiano Think of it more as the difference between > (greater than) and >= (greater than or equal).
@liketowrite, you probably mention my post (page 275, post #4125 on 01/15/2016), this link
http://talk.collegeconfidential.com/sat-act-tests-test-preparation/1816872-psat-discussion-thread-2015-p275.html
@Ynotgo, you are right. Bravo
I like your rationale for splitting at 217-218.
If this corrects the accuracy of the table then that leaves CA with a 217 nmsf cutoff and @thshadow wins the prediction contest. (and my dd wins a college education)
@Mamelot Nothing using concorded tables. Just a simple still stuck on the SI percentile definition and why are there such large numbers of high scores showing up in schools.
No one has yet verified that what the SI percentiles mean.
If the SI% is based on actual test takers, then of the 1.7MM, 17k are at 99%, 34k by 98% and 51k by 97% with SI’s of about 209, 203, and 201 with a commended of 201. Everyone is happy as lower cutoffs should happen.
From the compass article pages 7-8: http://www.■■■■■■■■■■■■■■■/wp-content/uploads/2016/01/Problems-with-the-New-PSAT-Compass-1-26-2016.pdf
From page 8: It states 3% of students are reporting percentiles in the 99% range. We all know this cannot happen. The numbers being reported seem to be inflated as the article suggests (the change in percentile definition along with the 3MM National number). There are numerous posts about schools with 2-3 times higher scores that in past years (Walton, OK, MN)
From page 8: If a student who really should be ranked in the 97% and is showing up at 99%, then their SI that should have been 200-201 is now 205-213. So a kid in WY who should be a 201SI (SF at cutoff /Commended) is now a 209 (SF at cutoff /Commended) kid in WY.
Another very simple way to look at it is… If the kids at Walton have gone up in record numbers and the cutoffs are rising (except for the highest end), then the kids at the lowest cutoff states also go up in record numbers so the commended number also will rise.
@likestowrite I like your new percentiles. They seem to correspond somewhat to @thshadow, I think.
So, the problem is the percentiles that have multiple SIs, right? If 205 is 99, with the percentile creep, we know it’s now 98. Could 98 go all the way up to 208? I would guess your percentile creep would be less pronounced at the top end because there won’t be as many students having each of those scores - 228, 226, etc. The rational for your break-off appears reasonable.
@liketowrite, I understand your way. But how do you justify move up lowest 99+% to 3 spots up, at the same time you keep the lowest 99% the same.
Make it clear, we have 9 slots for 99+, and 5 slots for 99 (so in the 99 we have total 14 slots - don’t care + or not). When u say move up 3 slots, that mean you move 2 or 3 slots up with reference point still on the same reference point
It’s a necessary and sufficient condition that all scores go up so you can move 3 slots up?. I confuse myself to. Sorry, my bad
@liketowrite.
Another problems as mentioned by @michgeaux, at very top end or perfect scorer that mean that person usually ranked somewhere b/w 99.86% and 99.92%, where does the guy/gal go? One can not out-beat oneself?
I believe CB change scoring policies at least twice (2005, 2015 and may be 2009). There were many cons at the time, but at the end people accept their way and go with wind
@Shelt29 with #1912
“Another very simple way to look at it is… If the kids at Walton have gone up in record numbers and the cutoffs are rising (except for the highest end), then the kids at the lowest cutoff states also go up in record numbers so the commended number also will rise.”
God bless USA. We are smarter b/c we have new ruler that make us taller??? But we still have 34000 commended, 1000 NMSF and 15000 NMF
Re post #1897 - @likestowrite @SLParent @dallaspiano @thshadow and @Mamelot & @DoyleB and others on the “data team” – the table with percentiles - this report about 2014 scores can be helpful to see where states fall percentile wise typically for states based on percentiles - see pg 3.
As an example - NY / Texas & a couple other states typically have cut off SI’s of 218-219 - 99%ile – 4-5th from the top of the 11 slots in the 2014 report - so maybe based on the new chart in post 1897 that is a 216 or 217 - 217 is what TestMasters is predicting. Can we add to the chart in post 1897 and put estimates for each state next to the score level or a range based upon the percentile the previous cut offs fell in??
Do some of us miss @Plotinus – was so helpful too!
Thanks to all crunch the numbers and creating our own SI concordance!! Pretty impressive.
I’ve mostly lurked on this fascinating thread as I’m not a data cruncher like so many of you! However, I have been reflecting on all the various posts. My gut tells me the CB and NMC do not want a big controversy when NMSFs are announced. As such, it is simplest to keep this year’s SIs and the corresponding percentiles similar to the relationship of SIs and percentiles in previous years.
I researched prior PSAT/NMSQT threads and found a listing created in December 2011 by @clapuma. He/she compiled a list of the SIs range for each percentile for several years of tests. I’ve pasted the information below. I’ve updated the information with the data for 2012 and 2013 and also added the commended scores I could locate. A review of the data shows the range of SIs pertaining to a particular percentile are fairly consistent from year to year.
When comparing Missouri’s cutoff scores since 2008 to the below data, I can see that our cutoff has consistently been in the 98th percentile.
My daughter’s 210 SI concords to a 211 for the 2014 and earlier tests (using her 730V and 640M) and to a range of 209-215 (using her 36R, 37W, 32M). Her 210 shows as an upper 99%tile on this year’s SI chart (the notorious page 11!). If I discount for the “percentile definition change” inflation it perhaps becomes an upper or mid 98%tile. As a result, I am cautiously optimistic.
Perhaps I’m crazy, but I prefer to have some hope that the folks at the CB aren’t complete fools and have tried to maintain some semblance of consistency as they set curves, created the preliminary concordance table, reviewed their research sample data, etc.
I hope the below historical data will be helpful to the data gurus!! I look forward to reading the continued analysis as we await more news from the CB and NMC. Best of luck to all of you!!
TEST YEAR/GRADUATION YEAR
2003/2005
99+ 224-240
99 214-223
98 208-213
97 204-207
96 200-203
2004/2006
99+ 223-240
99 213-222
98 207-212
97 203-206
96 199-202
2005/2007
99+ 225-240
99 215-224
98 209-214
97 204-208
96 201-203
2006/2008
99+ 223-240
99 212-222
98 206-211
97 201-205
96 198-200
2007/2009
99+ 222-240
99 212-221
98 206-211
97 201-205
96 198-200
2008/2010
99+ 223-240
99 212-222
98 206-211
97 202-205
96 198-201
Commended: 201
2009/2011
99+ 222-240
99 212-221
98 206-211
97 201-205
96 197-200
Commended: 201
2010/2012
99+ 224-240
99 214-223
98 207-213
97 203-206
96 199-202
Commended: 202
2011/2013
99+ 222-240
99 211-221
98 205-210
97 201-204
96 197-200
Commended: 200
2012/2014
99+ 225-240
99 215-224
98 208-214
97 203-207
96 200-202
Commended: 203
http://phs.princetonk12.org/guidance/Spotlight/S03A7EC5E-03A7ECF9.1/understanding-psat-nmsqt-scores.pdf
2013/2015
99+ 224-240
99 213-223
98 206-212
97 202-205
96 198-201
Commended: 201
https://kleinhs.kleinisd.net/users/0012/docs/14-15_NewsInfo/PSAT_UnderstandingScores2014.pdf