New results for AP CS in Georgia, with puzzles

I mentioned a few weeks ago that the 2010 AP CS results are out.  Our external evaluator, Tom McKlin of The Findings Group, just sent us his report on trends in Georgia (and comparisons regionally and nationally) for AP CS, since “Georgia Computes!” started.  The results are positive, but include some fascinating puzzles.  Here’s his summary:

1.  The total number of schools passing the AP CS audit is down from previous years, but the total number of AP CS students is up.

2.  Georgia has more female AP CS test takers than any previous year going back to 2003.

3.  Georgia has experienced more dramatic gains in the number of students taking the AP CS exam than any other southern state.

4.  There has been a steady increase in the number of women taking AP CS exams nationwide.  I would expect that the Taulbee numbers should start to reflect that trend.

5.  Students who take the AP CS exam are ten times more likely to major in an area related to CS.

6.  Florida, Alabama, and Tennessee made dramatic gains in the number of women taking the AP CS A exam.  What is happening in these states?

I won’t address all of these points, nor even all of the puzzles in these data (e.g., Tom’s question in point #6). Let me get the bragging out of the way, so that I can get to the puzzle that I’m interested in.  Check out these overall numbers (just AP CS Level A, to be completely honest).  “Georgia Computes!” started in 2006.  I think one can see an effect there.

That’s a pretty amazing rise in Georgia exam numbers.  (Notice that our CAITE BPC Alliance colleagues in Massachusetts have reason to be pleased, too!)

Here’s one of the puzzles that I find interesting: After two years of essentially flat Black student performance in 2008 and 2009 (e.g., only about 10% of Black students who took the AP CS Level A test passed), we had a sudden jump this last year.  We had a drop in Hispanic student pass rate, but in keeping with historical trends.

But here’s the puzzle that I’m really struck by: With fewer classes, we’re getting more students taking the test!  If you simply divide the number of test-takers by the number of classes that pass the College Board’s audit for AP CS nationally, you’ll see that the yield of test-takers is low.  What could explain these results in Georgia?

I’ve been discussing these results with some of the folks here at the CECC meeting with me.  The best explanation that I’ve heard yet for both results is simply maturation.  Maybe this kind of effort (of Barb and her Institute for Computing Education and all of “Georgia Computes!”) takes time to develop to the point of measurable results (e.g., for the teachers to learn how to teach AP CS in the classroom, not just in the workshops).  Eventually, each teacher in each classroom becomes more effective and more productive, resulting in more students taking the test and more students passing.  It’s important for us to tease out what’s going on here, because if we could replicate it, it might offer a margin of safety for the CS10K effort.  We could still have enough test-takers for the new AP CS:P exam, even without 10K teachers in 2015, if each teacher has higher test-taking yield than we’ve seen in the past.

Best hope for CS Teacher Education is in-service, not pre-service

I mentioned a few weeks ago that I was very interested in high school computer science teacher education. I have probably talked to faculty at a dozen schools of education now.  My conclusion is that the best hope for developing more high school computer science teachers in the US is in-service, not pre-service. In education-speak, “pre-service” is before a teacher starts in the classroom, typically in undergraduate.  “In-service” teacher education might mean workshops or summer classes or on-line — it’s teacher professional development while a teacher is in the classroom.

I found only one pre-service high school CS teacher education (TE) program.  (More might exist in the US.)  In 14 years, they have had exactly 7 CS graduates. That’s a problem that I hadn’t thought about, but is obvious in hindsight.  Why should teachers want to become CS teachers?  Lijun Ni is finding a laundry list of challenges to becoming a high school CS teacher.  (Latest one that hadn’t occurred to me: A calculus teacher might take in-service TE to teach calculus better, but rarely to learn new calculus.  CS teachers have to learn a new programming language to teach every few years.  Which one is more attractive to someone who wants to focus on teaching?)  If we built it, would they come?

My best informant was David Jackson (who was my TA for educational philosophy a few decades ago) who runs a science TE program at U. Georgia. He says that the number one biggest problem that any TE program has is maintaining the relationships with teachers, schools, and principals for placing students in practicum and student teaching experiences. That raises a huge chicken-and-the-egg problem: Where are we going to get the classes to host student teachers in CS? As I understand it, the No Child Left Behind (NCLB) act requires teachers to teach subjects that they are “highly qualified” for.  To become highly-qualified pre-service, you must student teach in that subject.  However, in-service, you can learn a new subject and take a new certificate, and you don’t have to do more student teaching.  David also told me that a “minor” in CS wouldn’t work in his program.  The pre-service science education program has zero elective hours in 4.5 years. There’s no room for even a single class in CS.

Our best shot for increasing the number of high school CS teachers in the US will be through in-service development activities, like Georgia’s endorsement (a certification that a teacher can earn after earning an “initial certification” as a teacher).  In-service TE has some particular challenges. The main one is that your students are full-time professionals.  You might get them face-to-face in summers, or on weekends or evenings, but they are busy.  If we are going to develop new high school CS teachers, we have to provide ways of learning computer science that involves few multi-hour marathon sessions at the keyboard.  Our main method of having students learn computer science is through a form of apprenticeship: sweat at the keyboard, and we’ll tell you if you got it right.  That’s hard for full-time professionals.  Can we do better, or at least, differently?

Oral History of CS Videos

Posted to SIGCSE-members —  a great finish to CSed Week.

CS Ed Week – take a look at the newly launched YouTube CEOHP channel

( http://www.youtube.com/user/CEOHP )

containing short video interviews with Computing Educators. This is part
of the Computing Educators Oral History Project whose newly revamped
website launched today ( http://ceohp.org ) .

This NSF-funded project will be archived and hosted by the Charles
Babbage Institute. It will continue to be under development with
curricular materials and additional interviews being added in the near
future.

Barbara Boucher Owens and Vicki Almstrum