The Dilution of the Perfect ACT Score

You can almost stop reading now. Perfect ACT scores were 10 times more common… an order of magnitude more common!… in 2018 than in 2001. This is after adjusting for the rising number of people taking the ACT. Data from ACT, analysis and snazzy Excel graph by the author.

When I was a junior in high school in 1982, I had a good day on the American College Testing (ACT) standardized test, one of the gateway tests for college admission. I did a little more prep than my friends. I bought a book and practiced a little bit with it. That was unusual in my middle-class public high school in Birmingham, Alabama. There were no test-prep classes that I knew of. No coaches. No online help; this was 1982. You just got in the car and drove to the testing site and took the test and went home and waited weeks for the score.

I remember being called to our high-school senior counselor, and found out the results. I made a 34 composite score. Back then, the highest score you could get was a 35, not 36 as it is today. And the only way to do that was to get a perfect-on-all-sections composite of 34.5, which was rounded up to a 35. I remember looking at the ACT report in my counselor’s office, a 10% sample of all test-takers, that suggested (by my memory) that only 8 people in the nation made a perfect score. I wasn’t one of them, but I still had had a good day.

That was the Eighties. And people of my generation and older, especially in the South and the Midwest were the ACT has been most popular, remember the ACT that way: as a test that very, very few people ever made a perfect score on.

But things have changed.

In the 2010s, there have been more and more stories coming out of school districts about perfect-ACT scores. (I’ll leave the SAT out of this story.) Schools love these stories for PR; journalists love these stories, they fill a happy-news hole; parents love the stories; it’s a win-win-win. And it fits in with the test-score/college-admissions mania that grips the United States. These news accounts are not unlike coverage of lottery winners.

But when 20 students at one high school in Ohio all make a perfect 36 on the ACT, then you might wonder if the perfect ACT is as rare and newsworthy as it used to be. It would be a good question to ask. I’ll try to answer it here, succinctly.

This story is in no way intended to malign the intellect, ability, achievement, etc. of students who make the top score on the ACT. I’m not jealous of them. This is not personal. As a parent of a student who also had a very good day on the ACT (and was embarrassed at the attention he got for it), as a member of a public school board member, and as a consumer of education news, I got interested in this subject. And, as a scientist, I’m used to finding and crunching numbers.

I’ve compiled, largely from ACT’s own annual Profile Reports, data on the number of perfect (36) composite test scores on the ACT from 2001 through 2018. Here’s the telling graph, done in minimalist Excel format:

The number of perfect composite scores (y-axis) on the ACT for each year of the 21st century (x-axis, leading digits of year omitted).

That’s a steep upward climb. In 2001, the number was 89. By 2013, it crossed over into four digits to 1,162. In 2018, it was 3,741. There’s no real cherry-picking issue here; in 2000 the number of perfect scores was 131, but that doesn’t change the arc of the overall curve. The number of perfect ACT scores in 2018 was 30–40 TIMES more than circa 2000. (Data for 2019 are not yet compiled, to my knowledge.)

I can hear you saying, “But, but… more students are taking the ACT! You have to account for that!” And so I will.

In 2000 and 2001, just over 1 million high school students took the ACT. By 2018, that number had risen to a little over 2 million. (ACT and SAT have been in a market-share war, with ACT taking the upper hand and then SAT responding with changes to its test to regain market share, a story told well by Paul Tough in this book.)

In the graphic below, I eliminate the number of test-takers as a factor by plotting the percentage of perfect scores as a ratio of the percentage of perfect scores in 2001. That way, if the curve if flat, you know that the rise in number of perfect scores in the figure above is just due to population increases plus the ACT getting more market share. If the curve rises with time, however, you know that something’s going on that is making the same number of test-takers have a higher occurrence of perfect scores in the 2010s than in the 2000s.

And, the answer is:

Ratio of perfect-ACT-score percentage (vs. all test-takers) to the perfect score percentage in 2001. The value in 2001 is 1. The x-axis is year, from 2001 to 2018.

So, in 2018 the occurrence of a perfect ACT score was TEN TIMES MORE LIKELY than in 2001. And in 2016, too.

Furthermore, the slope is pretty steep. In 2018, the occurrence of a perfect ACT score was 4.5 times more likely than in 2008.

Everyone will have different explanations for this. I’m more interested in just putting the data out there. Some explanations, not mutually exclusive, could include:

  • Top-end students (defined by test scores only, with all the biases inherent) in 2018 are smarter than top-end students in 2008, and both groups were smarter than top-end students in 2001;
  • Top-end students are getting more test-prep help each year;
  • Top-end students are better-practiced at standardized tests now than in the past;
  • The ACT is easily “gamed” somehow by top-end students;
  • Students are taking and re-taking the ACT more and more, permitting top-end students to eventually make perfect scores.
  • Other explanations.

Your explanations may be different, and I welcome them.

A geography professor and meteorologist at UGA in Athens, GA. I write about news, sports, weather, climate, education, journalism, religion, poetry, the South.