A classmate keeps lamenting that the SAT has become much easier today than it was 50 years ago, but I haven’t been able to come up with any reliable statistics.
In particular, he says that an 800 score on either the verbal or math test was extremely rare in the 60s & 70s, but now are relatively common for applicants to many of the elite universities. Is that really the case? does anyone know the prevalence of 800 scores then and now?
Higher scores don’t necessarily mean that the test is easier. The difference in scores could be due to better education.
It is much more common these days to link funding to test scores, so there’s also the possible correlation that scores are better because education is much more focused towards the tests (which may actually mean a worse education).
That said, I did not find the prevalence of 800 SAT scores, but I did find a site listing the average SAT scores from 1972 to 2019.
Math scores started at 509, dropped into the 490s during the 1980s, slowly rose to peak at about 520 in 2005, then dropped back to 508 by 2016.
Critical Reading scores started at their highest, at 530 in 1972, then slowly dropped to 494 by 2016.
From that, it’s pretty clear that the tests were not getting easier in the range of 1972 to 2016.
In 2016 the tests were changed. A Writing section was added in 2006, and Writing and Critical Reading were combined in 2016. Scores since 2016 show a significant and immediate jump by about 25 points in math and 35 points or so in Reading and Writing.
Since it is not likely that the average student changed so significantly in a single year, this most likely indicates that the average student does score higher on SATs since 2016.
Statistics from here: Average SAT Scores Over Time: 1972–2022
Note - this is an SAT prep site. I don’t know where they got their numbers from.
The SAT was entirely reworked in 2016, but that wasn’t the only time. There was a massive rework in 2005, as well, and a re-calibration in 1996. You just can’t compare scores over time. Even the testing scale has fluctuated (it was 2400 points from 2005-2016. Actually, it was 3 separate scores that were never actually added up, like 3 tests. But everyone informally added them).
As a person who deals with advanced academics and college access to elite schools, I would say that they SAT is a BETTER test now than it has ever been: it’s a much closer (though still very imperfect) assessment of the skill set you actually need to have in college. No weird analogies, no obscure vocabulary, no pretending there’s this abstract quality called “aptitude” that can be teased out and assessed separately from knowledge and skill. It also does not rely QUITE as heavily on shorting time in order to create a nice bell curve. Is it easier? I don’t even know how to answer the question.
In terms of students, I think we do a MUCH better job of educating the most highly motivated/talented/engaged learners than we did a generation ago. Now, some see this as a wonderful improvement in American education and others as an example of accelerating the rat race. But if a kid wants to or is forced to grind, there exists several clear pathways to a much more advanced education than was widely available in 2000. Furthermore, explicit, targeted test prep is much more of a thing. Good SAT tutors can command $200/hr and, in some areas, never hurt for business. Poor SAT tutors can charge $75. So yes, I think it’s quite possible that the kids are just better. On the other hand, there’s a pretty compelling counterargument that they are still about the same by the end of undergraduate studies: that early edge isn’t always capitalized on.
Another note: who TAKES the test has changed significantly. Last year, about 2/3 of high school students took the SAT-- and there are whole states where everyone takes the ACT. So most kids are taking it. In 1975, only about a thirdof graduating students were taking it–the ones who knew they were college bound. So the cohorts have also drastically shifted, making comparisons even more difficult.
The number of 800s isn’t really a good metric anyway. At best, it looks like the average score has gone up 30-40 points. That’s at best. Out at the high end, we’re still a few standard deviations above the mean.
What is true is there are more students taking the test. More students = more students getting 800s, even if the percentage of total students getting them hasn’t changed. And there are now plenty of programs to help students take the tests.
There are more high school QBs throwing touchdowns now than in the 60s or 70s, too. That doesn’t mean we’ve somehow made throwing a football easier to do.
The Flynn effect may or may not have something to do with that.
This doesn’t tell us anything useful though. As mentioned, the test isn’t static over time.
More students taking the test would probably lead to both a greater number of perfect scores, and a lower average score (assuming all else were equal). The people who wouldn’t have taken it in the past had some reason for not taking it, and those reasons probably correlate fairly well with being less academically skilled. If they’re now being pressured into taking it anyway, they’ll likely get low scores, and hence drag down the average. Though of course, there will also be some who would have had other reasons for not taking it (lack of finances, low self-esteem leading to an underestimation of ability, etc.) who would nonetheless do well on it, and those students are extra opportunities for perfect scores.
Another thing that’s happened in recent decades is that schools (at least, good schools, which are the ones most likely to produce perfect scores to begin with) have started providing a lot more opportunities for their top-performing students. When I was in high school in the 90s (at a private, exclusive school), my school offered three AP courses, and I took all three of them. Nowadays, at the public (though suburban, so better than the inner-city schools) schools that I teach at, there are students taking eight or nine AP courses, and could take more if there were room in their schedules for them. It’s debatable whether education has improved for the bulk of the student body, but for the top performers, it absolutely has.
Correction: the Writing section was added in 2005, but revamped significantly in 2016 (they basically stole the ACT’s format lock stock & smoking barrel).
As others have pointed out, you can’t really compare scores across eras. Not just the test format changes, but the testing pools do too.
2015 changes [note I’ve been tutoring for both tests for 20 years now]:
Writing probably got a bit easier. The 2005-2015 question types were more than a bit twee.
Reading however got significantly harder. Welp, how do you do that? You make the distinctions between answer choices very fine, often hinging on the connotations of specific key words. You’ll have “trap” answers which would seem to be pretty close matches-except another answer is a teeny-tiny bit better & more accurate match for what the author said & what the question specifically wants. I tell my students that they have to be rigorously logical, and sloppy thinking will sink them.
The issue there is that it will be harder to make the Reading more difficult going forward w/o also making it more unfair, where you basically have a coin flip [what I tell my students is a 50:50 guess] and that’s all you can do, bubble in, move on. This has proven frustrating for a lot of my students.
Math has had a significant revamp, and in my professional opinion is definitely harder. While all of those abstract “puzzle” like questions have gone bye bye (incl. the dreaded Defined Operations ones) in favor of concrete questions dealing with real world concerns, there is a much greater emphasis on higher level algebra [systems, quadratics, coordinate graphs] as well as statistics, which many of my students have had little to no exposure to. It is riddled with convoluted multi-step “timewasters” where many students have no idea how to get started.
When they ask, I basically tell my students that they would be amazed at how easy a test from say 2000 would be (unless your vocab skills weren’t up to snuff).
But now we have a curious phenomenon, where specific editions of the test can be significantly harder or easier. June 2018 & August 2019 were seen as much easier than normal. Why is this a problem? Because the scoring curve becomes “harsher”, where missing 5 questions means you only got a 680, vs. a 720, meaning you can’t really afford any careless mistakes.
The classmate is talking about something very specific, which is that it’s easier to get an 800 score than it was 50 years ago. This is definitely true. In 1995, the test was re-centered, so that a 1994 score of 500 verbal was now 580 in 1995. A 1994 score of 500 math was now 520. Every pre-1994 score between 730 and 800 verbal was now 800. Every pre-1994 score of 780 to 800 math was now 800.
I think this hurt the candidate who tested exceptionally well. I remember reading in one of the early college admission counseling books written before 1994 that a score of 750 or higher on verbal made the elite colleges sit up and take notice, because it was extremely rare. But now it’s not, and the school doesn’t have a great way of distinguishing the very, very top students from the test score perspective.
Other changes may have had influence at the top end as well besides the re-centering. The number of students who have gotten high scores has gone up over time since 1995. At least 25% of the students at some of the most elite schools got 800 on math. The 25th percentile of scores on math for MIT and Caltech is 790, which means 75% of their students got at least 790 in math. That would not have been true in the 1980s, and I don’t think that’s purely explained by re-centering. SAT tutoring is widespread now. As has been mentioned, there is a big emphasis on AP exams for those wanting to go to top schools. The percent of high school students who take calculus more than doubled between 1990 and 2009 (7% to 16%), with increases in the percent taking trig, geometry, etc. I don’t know stats for English courses, but I would guess there might be a push in difficulty on that side as well.
Just changing the scaling of the scores doesn’t change the test itself. You’re conflating two distinct things.
There are just too many variables here to be able to say the test is “easier” now than it was 50 years ago, if you’re determining that simply by the scaled scored. How can your friend say that it’s “easier” to get an 800 score now than it was back then? Did he go back in time to take the test? Just because more people are scoring 800 doesn’t necessarily mean that the test itself is easier.
At some level, you can’t separate out the test from the environment it’s in. One could say that the test is just as hard as it’s always been, but more tutors means more 800s. Alternately, one could say that the tutors are making the test easier. I’m not sure that either view is inherently more valid.
The recentering is irrelevant. The test has been completely revamped twice since then.
In 19851 in 110,000 takers got a perfect score. In 1990 1 in 120,000 takers got a perfect score. After rescoring in 1995 1 in 1,400 takers got a perfect score. In 2005 after the writing section was added 1 in 2,400 takers got a perfect score. In 2006 1 in 5,000 takers got a perfect score.
From that it seems like the SAT has gotten a lot easier to get a perfect score on. Some of the increase may be the changing demographics of test takers. The percentage of asian test takers has gone up to about 10% of all takers. Most likely explanation is the test has gotten easier.
In what way is this a relevant piece of information?
Asians do significantly better on the SAT than other ethnicities and so a higher percentage of them taking the test would result in more perfect scores.
OK, how much better do they do than other ethnicities? Is this higher performance consistent for different numbers of Asians taking the test? If the proportion of Asians taking the test increased to 10%, what was it before?
You need a lot more information there for your statement to be meaningful, unless your only intention for it was racism.
Asian average score is 1223, while overall average score is 1068. that is one standard deviation better. If my math is right and the standard deviations are similar between groups, that would mean 2.5% of Asian test takers would have a perfect score and .15% of overall test takers would have a perfect score.
In 1981 3.5% of test takers were Asian, and in 1990 7.7% were Asian.
Your math is not right. There is no way to determine the proportion who got a perfect score from the average score. And you still haven’t answered whether the Asian average score has remained the same while the proportion of Asians in the population increased.
It means it’s easier to get a perfect score, not necessarily that the test itself is easier.
Any test with a writing component is subject to norming variation.
