Based on what I heard, October tests are usually harder than other tests because it is the last test for seniors. I took my first exam on October and yes it was definitely harder. But I’m not so sure if it’s legit tho
File it under “it is what it is.”
No, I don’t think they are designed to be harder. If it turns out that it is, the scores will likely be normalized.
I would not choose test dates based on “what I heard.” Good luck.
No, it’s not true.
Why would College Board do that?
Because they use different questions for each sitting of the test, there will always be people who think one version is harder than another.,typically the one you just took was undoubtedly the hardest ever given.
Complete urban legend with zero factual basis, @kwcho19
No. Harder than what? From posted anecdotes, last fall beginning with August, most or all of the test sittings had reading sections that were apparently a bit harder (one noticeably more difficult passage) than the published practice tests and that seemed consistent through May of this year. June may have been an anomaly. The New test has only been around for two years and there are no patterns in difficulty for certain months. The only occurrence that has happened twice so far is an especially large number of different test forms given during the June administration; a few rare forms from June 2017 turned up as administrations last fall.
The SAT is also offered in Nov and Dec, which is not too late for RD apps.
(Also worth noting that (A) there were no patterns of difficulty on the old SAT, which had a much larger sample size, and (B) for all its many faults, the College Board is very, very good at equating its tests, so harder tests have gentler curves and easier tests have harsher curves. In my 17 years of experience, there are no “easier tests” when curving is factored in, and there are absolutely no “easier” or “harder” months.
I agree: no such thing as a hard or easy month.
But @marvin100 – I don’t know how you can know if the College Board is good or bad at equating. Do we have access to data to support or refute that?
One thing we do know from looking at the old SAT and its curves (which Erik the Red kindly posted here: https://www.erikthered.com/tutor/SAT-Released-Test-Curves.pdf) is that they were very good at writing tests that would have consistent curves. Pick a raw score and scan through the tests. Most of the scores are within 10 points of the average, with a handful of 20 point deviations and an occasional-but-rare 30 point outlier.
As for the June test:
It may be that they are having trouble writing exams with that same consistent difficulty level. Or maybe we just observed an outlier. Or maybe the equating process itself is more difficult for them now that they don’t have a blind equating section.
Also, one more cynical note: maybe they were never as good as it seemed at writing tests. We don’t get to see the raw data and calculations that go into creating the curve. If they really wrote a clunker, would they produce the truly honest curve? Wouldn’t that be an admission? Maybe that is what June’s curve represents.
@pckeller It is believed that the New test is written in-house at College Board whereas the old test writing was subcontracted to ETS for decades. There was some turmoil within CB over the new test being ramped up too quickly; Reuters had some articles a while back. ETS still has some role though it is unclear.
@pckeller - yeah, Erik the Red’s repository is pretty strong evidence for their equating ability (we may be talking past each other–I’m including curves in “equating”), but also it’s easy to see how consistent most people’s scores are across multiple tests.
Now I’m not saying they’re great at making tests, just that they’re good at equating them (with statistically effective curves). The post-2016 exams, in particular, have been uneven, as they rushed out the reboot and have been wobbling towards an equilibrium ever since.
I’m more bothered by the gruesome and misleading flaws in the “CB” Khan Academy practice material, though, tbh.
@marvin100 could you elaborate on “the gruesome and misleading flaws in the “BC” khan academy practice material”? My DC is using it to prepare for his upcoming test. Thank you.
The M and W sections are the worst, but they’re not my specialty. In R, though, Khan’s material features two question types that no longer come out on the real exam: analogy questions (“Which of the following situations is most similar to that of lines 24-32?”) and EXCEPT questions (“The author mentions all of the following EXCEPT”).
@makemesmart - I’d still use those tests, as they’re better than most of what’s available, but I’m for holding the CB’s feet to the fire on these exams that they claim they’re vetting.
The college board want to help people get college places. Why would they make it harder to get a good score as people want college places
@pckeller “Or maybe the equating process itself is more difficult for them now that they don’t have a blind equating section”
What is a blind equating section?
I have done extensive prep through Khan Academy and have yet to see the type of question @marvin100 is mentioning. I haven’t seen any glaring aberrations in the math or writing, either.
@suzyQ7 I was referring to the section on the old SAT that did not count toward your scores, with no way for students to know that they were working on that section. That’s where they could bury the questions that were used for equating and also where they could pre-test questions for possible future use. Kids had to try their best because it was very risky to assume you were in an equating section when you could easily be wrong. So it generated lots of useful data for them. I don’t know how they are getting the data they need to do these tasks now. I know that they make some kids take extra sections (if they are not writing the essay) and they are intentionally vague about whether those sections count. But that might not be working as well as the old equating section did.
@awesomepolyglot - keep going. There are many analogy q’s and even except q’s (so far there’s only been a single analogy q on a real test since the revision).