A first look at the impact of AP Computer Science Principles

The new Advanced Placement Computer Science Principles class is making a positive impact on both gender and racial/ethnic diversity in computer science.

The College Board, the organization that administers Advanced Placement (AP) classes and exams, as well as the Standardized Achievement Test (SAT), has made an effort to increase diversity in computer science (CS). The traditional CS course, AP CS “A” (CSA), was being taken by a low number of girls and racial/ethnic minorities. (For a while there was a more-advanced class, AP CS “B,” which had even worse diversity, but that class was discontinued.) So they introduced a more introductory-level class, AP CS “Principles” (CSP), aimed at providing the basic principles students would need for understanding computer science. The hope was that this would provide more equitable access to students who hadn’t previously had as much exposure to computers and computing principles. Indeed, the CSP class saw much-higher enrollment of female and minority high school students. (I previously blogged about the gender balance of the AP trends in 2020 update: women in computer science; I didn’t find the racial/ethnic data.) Those of us who follow diversity in tech metrics have been waiting to see if that increase leads to increased enrollment in CSA and then increased rate of majoring in CS in college.

Well, the College Board released findings in December (findings summary | full report) concluding that the CSP class indeed increased female/minority enrollment in CSA, as well as increased the rate at which female/minority students enrolled in CS majors in college. They also found that female/minority students taking CSP were more likely to go on to take other AP STEM classes. Wahoo! That’s a lot of wins!

They did this study by trying to control for other variables: the students who take AP classes fill out a survey self-reporting their gender, ethnicity, parental education, and high school GPA (HSGPA). The College Board also had their SAT scores, and their declared majors at 4-year colleges. They did a comparison between students from the graduating class of 2016, before CSP was available, and the graduating class of 2019, who had chances to take CSP in grades 10-12. They matched up students from 2016 and 2019 who had the same or very similar results for all of these variables, in an attempt to get a comparable result on the rate of taking CSA and the rate of majoring in college CS, with and without CSP.

There are all sorts of numbers in the full report, but here are some choice conclusions from their data:

  • Female CSP students are 5 times more likely to major in college CS than their comparable female students from 2016 who never had the chance to take CSP.
  • Black CSP students more than twice as likely to major in college CS
  • Hispanic CSP students over three times as likely to major in college CS
  • First-generation students (those whose parents do not have college degrees) were over three times as likely to major in college CS.
  • Students who took CSP went on to college STEM majors at about a 10 percentage point increase compared to before CSP was offered. This was consistent across all genders and ethnicities.
  • Female and minority CSP students went on to take CSA at about double the rate compared to comparable students from 2016.

Remember, these comparisons were done by comparing the rate at which students from 2019 entered these classes/majors vs. the rate at which their comparable cohort of students from 2016 entered them. These are amazing conclusions!

I do have one complaint about these conclusions, though. They did not control for the difference between 2016 and 2019. That is, these conclusions assume that 2016 and 2019 had otherwise comparable forces acting on students to take CS in high school and college. But clearly that is not true… The evidence is that for the past decade, enrollment in both AP CS A and college CS has been growing strongly, even before CSP was introduced. See the graphs in my 2020 update: women in computer science post. CSA enrollment grew from 2010 to 2016 even without CSP being introduced. Computer science degrees-awarded have been growing since 2009. So, the students from the graduating class of 2019 would have been majoring in CS at a higher rate than those graduating in 2016 even without CSP. I have no doubt that CSP made a positive difference on both the diversity and number of students entering CS in college – but the impact of CSP was a little bit lower than what is stated in the report.

Still! Hooray for diversity in CS and STEM! These are numbers to celebrate! 🙌🎉🥂

And, I can’t wait to see the CS graduation rates for the college class of 2023. At the rate the government publishes data, I’ll have to wait until 2025 to see those numbers. I’ll be waiting…

Note: cover image courtesy of #WOCinTechChat, wocintechchat.com

Leave a Reply