ECEP Alliance: Measuring CS Education progress in US states

Posted by CSTA on behalf of The Expanding Computing Education Pathways (ECEP) Alliance

The Expanding Computing Education Pathways alliance ( is an NSF-funded alliance to broaden participation in computing. Our focus is on the education pathways (from elementary through high school, to community college and universities), because that’s our best chance to reach underserved populations. Our challenge is that education policies vary dramatically from state-to-state, so we can’t come up with one solution that works for everybody. A model that we promote for getting started in a state is:

  • Step 1: Find a leader(s): You need a leader (or a couple) who will take the initiative and who see(s) the big picture of how schools, higher education, businesses, and state politics have to work together to make change within a state.
  • Step 2: Understand your state’s policies: Who makes the decisions in your state about high school graduation requirements, teacher certification, and high school curricula? Where does computing fit within your state’s policies? Think about writing a landscape report that lays out the current state of computing education within your state. (There are several of these available at the CSTA website, such as one from South Carolina and another on Maryland.)
  • Step 3: Gather your allies: Efforts that speak with multiple voices from different sectors promoting computing education tend to get more influence in state government. Computing education summits are where you meet face-to-face, to talk about shared goals and come up with strategies that all the allies can work on.
  • Step 4: Get initial funding: Landscape reports, summits, and other meetings take some small pots of funding to get you started, before the big ticket items, like professional learning opportunities for all your high school teachers.

If every state has different policies, how do we measure progress? How can we tell that things are getting better, or which states are moving ahead and faster than others? Every year, Barbara Ericson of ECEP collates the College Board data on who took the Advanced Placement® Computer Science (AP CS) exam. Data on AP CS doesn’t cover all computing education in a state, but it’s likely a close measure and it gives us a way of comparing progress in states. The College Board doesn’t know how many AP CS teachers there are, but does track how many schools pass the audit which allows them to offer AP CS. Most schools that pass the audit have exactly one teacher, so counting schools that pass the audit is a rough count of AP CS teachers.

Barbara’s analysis of AP CS A in 2013 (available here) got a lot of press coverage, including the New York Times, CNN, Slate, and Washington Post (see a list here). Barbara has a preliminary set of results available now on the 2014 data (her analysis and data are available here). Here are some of the national highlights:

  • The number of AP CS exam takers rose 26.29% in 2014 (from 29,444 to 37,327).
  • The number of schools passing the audit rose a little over 10% (from 2,252 to 2,525). The big difference in those two statistics (27% more test-takers, only 10% more teachers) means that each teacher is getting more students to take the AP CS exam.
  • Women, Black (the College Board’s category), and Hispanic exam takers all increased about 35%. That’s faster than the overall exam taker growth at 27%, but just barely. In 2013, 18.5% of exam takers were female. In 2014, 20.0% of exam takers were female.
  • A smaller percentage of students passed (from 66.86% to 61% overall), and that was true within demographic groups, too. 62% of female exam takers passed in 2013, but only 57% this year. A drop in pass rates is not too unexpected if we are getting more students into the exam, especially if new students are coming from schools and teachers new to teaching AP CS.

When we get to the individual states, the picture is more complex, but is still striking in terms of how little AP CS there is yet in some states.

  • 18 states had less than 100 people total take the AP CS exam in 2014. Montana had only 4 exam takers (all male). Mississippi also had four exam takers (one female), and though the state is 38% Black, they had no Black AP CS exam takers. Wyoming didn’t have a single AP CS exam-taker in 2013 or 2014.
  • California leads the nation in number of AP CS exam takers and had the biggest gain in exam taking, with a 34% increase from 2013 to 2014. Florida jumped from 8th in the US to 4th with a 39%. Maryland had surprisingly little growth from 2013 (from 1629 students in 2013 to 1639 in 2014) and dropped from 5th to 8th.
  • California is also the most populous state. Maryland has the most exam takers for its population, followed by Virginia and New Jersey.
  • We are nowhere near gender-balance in AP CS exam taking. With 1/4 (25%), Mississippi has the highest percentage of females taking the AP CS exam. The next three top states are Washington (260/1048 = 25%), Oklahoma (42/171 = 25%), and Texas (1102/4551 = 24%). The states with the least female participation in AP CS exam-taking are Montana (0/4), Wyoming (0/0), Mississippi (1/4), North Dakota (1/14), Nebraska (2/71), Kansas (3/40), Alaska (4/30), South Dakota (4/29), Utah (5/104), and Delaware (7/79).
  • Barbara is still going through the race data, but even the bright spots still aren’t that bright. Maryland had the most Black exam takers (192) with a 30% pass rate, which means that 12% of their exam takers were Black. 30% of Maryland’s population is Black. Texas was second (161 exam takers, 40% pass rate), which is 4% Black exam takers in a state that is 12% Black.

These results are positive in terms of growth, but we have a long way to go. AP CS is smaller and more gender-skewed than any other AP exam (see Brian Danielak’s insightful visualization here). We use AP CS as a measure for CS education in the United States overall. Computer science in high schools is rare, mostly male, and mostly white or Asian. That’s what we’re trying to fix.