@desideratum The cutoff of 209 is for Commended status and is the same for all the states. The cutoff for NMSF for California will likely be somewhere in the range of 218-224, but, of course, we won’t know for sure until September.
I haven’t been following the latest cutoff discussions so closely, but I would like to make one observation.
People should not be surprised that this year’s commended cutoff is the SI that concords to last year’s commended cutoff. The commended cutoff is a percentile (97th percentile). The concorded scores are scores with the same percentile rank. So a 97th percentile SI last year (203) concords to a 97th percentile SI this year (209) and each 97th percentile SI is the commended cutoff for its respective year.
The only reason any of this is surprising is that the percentile tables released by CB were bogus.
I can’t remember what data is available for past years. Has the College Board released enough data in past years to create a chart that compares the state cutoffs to the national percentiles? If so, then we know that 209 equals somewhere around 3.1-3.3% in this year’s national percentiles (50,000 out of 1.5-1.6 mil qualified test takers). We also have an example of a possible curve for this year’s test from the Testmasters TX data. We could come up with some fairly accurate national percentiles from there.
This data exists, but it’s really hard to gather it all in one place. I spent a lot of time collecting and analyzing data for PA, and in the past, PA has landed roughly 8-9 points from the 99+ cut. However, a point now does not equal a point then, so it’s not all terribly useful today. But yes, bigger states do typically stick close to a similar yearly percentile. All bets are off though for smaller states.
The state reports don’t show the total scores, do they? When looking at NJ, I only see Reading, Math, Writing breakdowns and not total score for PSAT.
The state reports only show subject scores - which makes it really fun to construct total score/SI’s from that!
I have asked my S’s GC again for scores from his school (which produces around 10 NMSFs every year) and again she didn’t provide any info. I guess they are busy with helping seniors.
So here is a different way of hashing this: In 2014 the 202 Commended score was the bottom of the 97th percentile (https://■■■■■■/NLZ5xI). If we assume that the shape of the curve for the Testmasters TX data matches the shape of the curve for the nation, then this would be the breakdown of the score ranges if 97% starts with 209 as 202 did in 2014. I don’t know exactly how the College Board breaks down the borders between their percentile scores, so I’m leaving some leeway. I’m also assuming 99+ is the same thing as 99.5.
97% = 209 to 213 or 214
98% = 214 or 215 to 218 or 219
99% = 219 or 220 to 221 or 222
99+ = 222 or 223 to 228
The Percentiles for the 2014 test were:
97% = 202 to 205
98% = 206 to 212
99% = 213 to 223
99+ = 224 to 240
@candjsdad Thank you for this. In your extrapolation, the ** first ** 99+ percentile is 222. Comparing to the official Understanding scores (with the BOGUS percentiles) 222 is the ** 9th ** 99+ percentile.
https://collegereadiness.collegeboard.org/pdf/2015-psat-nmsqt-understanding-scores.pdf
Although my disdain and expectations for CB data is at an all time low, 9 places off from last year in the percentile chart seems to be ‘unrealistic’?
Editing my own observations. A 97 percentile for commended on the current report would be 200 or 201. Since we know commended is 209, indeed the bogus percentiles are 9 places higher than you’d expect. Further proving that CB’s horrible percentile report is 100% inaccurate and misleading and your analysis seems to be spot on.
When is CB supposed to republish the tables?
Fortunately(?) CB always published previous year (2013) percentiles and the PSAT 2013 commended was 201 so you’ll need to shift everything down a notch. I wish we had the 2014 percentiles - they would have been published this year in the Understanding Scores 2015 book. But instead we get the not-so-helpful “Research study” table.
Still, pretty scary. MN that year was a 215 which was the bottom quartile of the 99th. That would suggest a 218(?)* now which is four higher than last year’s 214. That’s a jump of 4 points for a mid-range state that doesn’t have any other extenuating circumstances (such as a massive shift over to SAT/PSAT).
*Adjusted.
“When is CB supposed to republish the tables?”
Guessing that’ll be in the Understanding Scores 2016 report.
ugggh…
Are we really sure that the shape of the curve would be the same nationally as for TX? I believe Testmasters conducted PSAT prep for the district, no? Wouldn’t the TX scores be slightly higher than a typical district elsewhere?
I think the 97 percentile comes from this calculation:
~50,000 students that move forward / ~1,600,000 valid participants in the contest = Top 3.12% of students that are commended.
@suzyQ7 I believe we concluded that those percentiles were not useful because, instead of being from the actual test takers, they were the percentiles of a separate experimental test group. That is why everyone gradually switched over to using the concordance tables to try to calibrate this year’s scores. For some reason, the concordance tables seem to be based on the actual test takers’ results instead of the experimental group. There has been much criticism of College Board for not publishing the actual percentiles from the real test takers.
Near the beginning of this long discussion, I calculated the concordance of some scores by pretending that someone got the exact same score on all three parts of the old test and then used the concordance tables to produce the new scores. From the results, I warned everyone that the top states’ cutoff scores might be a little lower, but that the bottom states’ cutoff scores would be higher.
@BunnyBlue do you know what entry that was of yours for the calculations? thanks.
@candjsdad what is your percentile estimates based on? Seems to me it doesn’t account for the compression. And most states wouldn’t have 99%ile cutoffs, which has never been the case.
@BunnyBlue The Commended scores are real scores from the 2015 test. The announcement that came to our house (we are homeschoolers) said, “With a Selection Index score of 209 or above, these high performers have shown outstanding academic potential. From approximately 1.5 million program entrants, each of these studens is among the 50,000 highest-scoring participants who will be recognized in the fall.” I think someone else called and confirmed that “approximately 1.5 million” was actually closer to 1.6 million. In either case, it is very close to 3% of students, and that is confirmed by looking at previous “Understanding Your Scores” reports.
@Mamelot True, I may be off a bit because the data in the 2014 “Understanding Your Scores” comes from the 2013 test. The big question, though, is whether the shape of the TX data curve is similar to the national curve, especially at the top of the range.