@mamelot That was my understanding of the scores above 220. There are about 325 in her grade. I don’t know how many testers but it was probably a lot. They all take the PSAT as sophomores but it’s optional for juniors. Most kids go to 4 year colleges so many probably do retake as juniors, especially with the new format this year.
@mamelot. I looked at the TX data in relation to NJ historical commended and NMSF numbers. If NJ kids score 2x the percentages of the TX kids at the very top (using test masters data) then the cutoff for NJ looks to be about 222 and the state would generate 2500 commended students. This assumes an accurate read of the testmasters chart and that there is a proportional relationship between the two groups. Basically I took the TX percent at each SSI and multiplied by two so the curve is the same, just higher.
Thanks @MomNJof2. Ugh. That anecdotal evidence speaks to a cutoff of 223 or 224. Double UGH. I was hoping for 221.
I’ve seen several times in this thread references to clumping at the even numbers. Does anyone know why or what causes it?
NJ SF’s are about .68% of total test takers and TX is about .82% so there is some variation there. Therefore it’s probably closer to 1.67x. Not sure. Does that push the cut-off up past 222? My concern would be that 224 is a hurdle. If you cut-off one lower you get too many SF’s so the cut-off would be 225 not 223 (or 222). I really hope that’s NOT the case. It doesn’t bode well for any state below NJ either.
@MomNJof2 Oops - the percentages should be the other way around! Sorry about that. .68% for Tx. and .82% for NJ. So the ratio should be 2.5x? Not sure that’s correct in any case.
@st8homom Fewer score combinations result in an odd SI than an even SI. To get an odd SI, a student needed to get 0.5 points on one section.
@itsgettingreal17 my ds had 2 wrong in reading, 1 wrong in writing, 0 wrong in math - 37,37,38 - si 224
@crazym0m Sorry, fixed it. That should have been odd SI. My D has an odd SI (219) because she got a 0.5 on the math section. Remember the formula was 2x CR + Math.
@itsgettingreal17 no biggie - also did you notice that there was not a single easy question on the writing section? Why have an “easy question” if you ask no easy questions???
There is also a big bump at 214-218 which DoyleB talks about around post #2185. Many had hypothesized about a multi-modal (not sure that was the word!) distribution this year.
@crazym0m. I hadn’t. My D did great on the writing and math and not so hot on the reading.
Did the test takers sort of show that the Writing overall was easier as missing 1 in Writing got you a 37 and missing 5 in Math got you a 36.5. If you missed 5 on Writing what did you get?
S missed three on writing and got a 35. There were only 44 questions for writing. 47 Qs for reading and 48Qs for math.
The writing curve was brutal… Yes, it was way too easy. I think -5 on writing gave you a 32.
@mamelot if the NJ to TX ratio is 2.5, then the cutoff is 223 but there are too many commended. All this assumes the curve is exactly the same ( just more of them at each level) which is dubious. But in the absence of any better data it’s all I have to crunch in excel!
@MomNJof2 …Class of 2016 PSAT cutoff was 225 for NJ…right? So you are thinking that it will be 225 again? Texas cutoff was a 220 for the class of 2016 and I wonder if Texas will be that high again? Testmasters was saying 219 for Texas would be the highest. What do you think?
@MomNJof2 your method seems pretty sound and you did about as well as can be expected given what we know! Anyway I just remembered (again from earlier conversations) that my percentages may be off because I used total program entrants not total test takers in the denominator. So who knows if the ratio that I gave you of 2.5x is even correct. And yeah, the curves will have some differences.
The thing that I’d look for is more evidence of a spike at 220, 222 or 224. Regardless of how the curve exactly looks, you know that the cut-off is going to be at the high-end and that a big spike could suggest a cut-off just to the north. (The same would apply to odd-numbered spikes of course but the chances are higher of even-numbered just due to the way that the scores were put together).
@tgirlfriend last year NJ cutoff was 225. I don’t think (or hope) it will be that high again but it sure looks on the basis of an unscientific sample that it will be in the 221-225 range. Or maybe the HS outperformed its history on the new test.
@mamelot I have been keeping an eye out for evidence on NJ scores but haven’t seen much yet. It sure would be helpful to have some real data. Way back one of the parents kept a log of self reported scores and they sure ran pretty high. But a limited data pool there too and probably skewed very high with CC readers.