The percentage of students taking (and passing) the AP Computer Science Exam has long been the benchmark used to measure strength in high school computer science education. And until the 2018 State of Computer Science Education report provided data on the percentage of high schools in a state providing any computer science course, it was the only measure.
But the percentage of students taking AP CS is just the top of the CAPE (Capacity – Access – Participation – Experience) pyramid. It shows how many students can pass a college-level CS course, but only after many inequities have already crept into the system. If we want to find and address these inequities, we need new metrics that measure what is going on in the lower layers of the CAPE pyramid.
Fortunately, the State of Computer Science Education report now uses the percentage of high schools providing any CS course in a state as its primary benchmark to measure a state’s strength in CS Education. The report also provides us with information on the percentage of students in a high school with a computer science course. And although the percentage of students enrolled in a foundational CS course was only available for 37 states in the 2021 report, this student percentage metric will provide even greater insights into the strengths and weaknesses of CS education in each state. With these additional metrics now available, this blog post will revisit the standard AP CS exam rate for what new information it provides.
This blog post will treat AP CS A and AP CS P the same – a student taking either exam is taking an “AP CS Exam.” The next blog post in the #CSEdAnalytics series will look at these two tests as separate entities.
Percentage of Students taking AP CS Exam
For the traditional benchmark, the national average for the percentage of students taking AP CS is 1.08%. The unweighted state mean is 0.88% (i.e., more populated states have higher rates) with a standard deviation of 0.60%. The following is a histogram of this percentage by state and a list of the leading states. Maryland is the runaway winner.
| State | Pct | Z-score |
|---|---|---|
| Maryland | 2.82% | 3.26 |
| New Jersey | 2.06% | 1.98 |
| Connecticut | 2.04% | 1.95 |
| District of Columbia | 2.04% | 1.95 |
AP CS/School Uptake Ratio
The AP CS/School Uptake Ratio indicates the percentage of students who attend a high school that offers computer science take an AP CS Exam. If students are not in a school that has a computer science course, it is unlikely that they will take an AP CS exam. The national average for the AP CS/School Uptake Ratio is 1.39%. The unweighted state mean is 1.11% (i.e., more populated states have higher percentages) with a standard deviation of 0.76%. The following is a histogram of this percentage by state and lists the leading states.
| State | Pct | Z-score |
|---|---|---|
| District of Columbia | 4.01% | 3.82 |
| Maryland | 2.93% | 2.39 |
| New Jersey | 2.35% | 1.63 |
| Connecticut | 2.29% | 1.56 |
The list of leading states by this metric reorders the leaders in the old standard metric. The District of Columbia is now the decisive leader. Another interesting factoid from this study is that even if all high schools in the State of Computer Science Education report offered computer science – an increase of 95% – we should only expect the number of AP CS exam participants to increase by 29%.
AP CS/FCS Uptake Ratio
The AP CS/School Uptake Ratio indicates the percentage of students enrolled in a (foundational) computer science course that took an AP CS Exam. Both AP CS A and AP CS P are (foundational) computer science courses that include at least 20 hours of programming instruction. Unfortunately, this ratio is only available for 37 states as only 37 states provided data on the number of high school students enrolled in computer science for the 2021 report. The national average for the AP CS/FCS Uptake Ratio is 22.4%. The unweighted state mean is 20.5% (i.e., more populated states have higher percentages) with a standard deviation of 12.5%. The following is a histogram of this percentage by state and lists the leading states.
| State | Pct | Z-score |
|---|---|---|
| Florida | 61.1% | 3.24 |
| Connecticut | 38.5% | 1.43 |
| Georgia | 37.5% | 1.36 |
| Virginia | 36.6% | 1.29 |
A very different group of leaders appears using this metric. Florida is now the decisive leader. The District of Columbia did not provide FCS enrollment data and did not appear anywhere on this list. Unlike the previous two metrics discussed above, a high AP CS/FCS Uptake Ratio is not necessarily good. As argued in the post Blacks in HS CS, too high an AP CS/FCS ratio indicates a lack of non-AP “pathway” computer sciences courses that underserved populations can take before taking an AP CS course. Even the national average of 22.4% is probably too high if one considers a similar percentage for AP Biology or AP Calculus.
AP Pass Rates
Although we will do a deeper dive into AP Pass Rates in the next #CSEdAnalytics blog post, I wanted to show data on AP Pass Rates to conclude this post. So far, we have looked at how the percentage of students taking the AP CS Exam changes from various perspectives. The AP Pass Rate shows the percentage of students passing the AP CS Exam with a score of 3 or better of those students who took the exam. The national average for the AP Pass Rate is 70.5%. The unweighted state mean is 67.3% (i.e., more populated states have higher percentages) with a standard deviation of 8.0%. The following is a histogram of this percentage by state and lists the leading states.
| State | Pct | Z-score |
|---|---|---|
| Montana | 78.8% | 1.25 |
| New Jersey | 78.1% | 1.36 |
| South Dakota | 78.0% | 1.35 |
| Connecticut | 77.3% | 1.25 |
While New Jersey and Connecticut appear among the leaders throughout this post, Montana – the leader in this metric – and South Dakota are new. The standard deviation between states is relatively much smaller than that of the other metrics above. Hopefully, the more “normal” distribution found in the AP Pass Rates is what happens to other HS CS education metrics after the metrics that show lower levels of the CAPE pyramid become better known and show weaknesses in each state’s approach.
Note that the AP Test Pass Rates listed above may differ from those reported elsewhere. As I will discuss in my next blog post that will look at AP CS A and AP CS P, I combined data from 2019 and 2020 due to the pandemic. Please visit the CSEd Analytics page for the underlying data and reports behind this blog post for more nuanced information. The attached reports will also show how your state compares to others in critical CSEd metrics.