Sunday, February 12, 2006

Advanced Placement

The CRA notes that while the number of students who take Advanced Placement exams has surged over the last few years, the number taking the Computer Science AP exams has dropped a bit, perhaps foreshadowing an even more dropping interest in undergraduate CS.

In many American high schools one can take AP courses that lead to standardized exams in a variety of topics that many universities will use to allow students to place out of some introductory courses. At least that was the purpose when I went to high school, but since then the AP exam has become a mainstay of the high school curriculum. Nearly a quarter of all high school students take at least one of 35 different AP exams. Student applying to good universities had better have several AP courses and exams on their record. Bush made AP exams a goal in his state of the union and Newsweek uses the AP test to rank high schools.

I have nothing against the AP exam in its original form, I took exams in math, physics and chemistry in high school and they saved me from some courses in college. But these exams have their drawbacks, as one has to teach to the exam. Gone in these course is the ability of teachers to experiment and students to excel in different ways.

We have this particular problem with the AP Computer Science A and AB exams. These exams force teaching in a specific language, currently Java, where teachers might have found other languages betters suited for presenting a variety of computer science concepts. The CS A exam focuses mostly on programming in Java, the CS AB exams does add some data structures and running-time analysis.

In high school (before the AP CS exam existed) I had a wonderful course that combined computer programming and probability. We don't see these kinds of interesting classes where the advanced classes in US high schools have to focus on exams.


  1. Perhaps you've answered your own question. There are less CS AP test takers because less people are able or interested in teaching it. (E.g. if they can't teach Python or Scheme the way they want, they'd rather return to being a system administrator and make more money.) Perhaps it's too early to tell, but this trend could predict more growth in the tech sector and not less.

    I think it's a big mistake to take AP courses. Unless it's a really large school, what gets taught as the intro in college covers much more interesting topics and in real depth.

    There is talk now of applying standardized tests to rate colleges. With any luck, the good schools like MIT will reject such testing and private institutions will not be subjected to them. (MIT is not a fan of the GRE either, and doesn't require it.) The Republicans used to favor school vouchers, which would have solved many of our educational problems. Now they seem only concerned about introducing as much regulation and problems as they can.

  2. I took the CS AP test the last year it was still in Pascal. My school didn't offer any CS, and all I knew was BASIC. But I thought it'd be fun to learn a new programming language for the purpose of taking the test. The test prep books are also what first introduced me to recursion and big-O notation. In short, I gained more from the AP test than from almost any of my actual classes. I think of that whenever someone suggests eliminating standardized tests in high school, in order to "liberate teachers' creativity." In many cases, that means liberating them to dream up new ways to stifle the students' creativity...

  3. First let me say that I am not responding to the question of whether or not the AP should equal college credit. Instead, I am responding to how it affects the high school experience.

    Personally, I benefited from AP classes because they introduced a level of rigor otherwise absent from my high school curriculum. They forced the teachers to present more challenging concepts. Teachers could no longer try to simply please the students or make the students feel like they were learning. Nor could they cater only to the lowest common denominator. They required the teachers to teach. Also, AP tests brought the students and teachers closer together as both felt like they were in a common battle against the test.

    That said, I can see how for some high schools, the level of the AP exams is not extraordinary, and for these schools they seem to serve no purpose. In extreme cases, they can actually prevent good teachers the freedom to teach optimally.

    This same good bad paradigm is true of many things (for example, no child left behind). If the school is already performing better than the level meant to be enforced, the added regulations make it no perform slightly worse. On the other hand, the new regulations (are meant to) encourage improvement schools which are subpar.

    Furthermore, the effectiveness of AP tests varies greatly by subject. For example, in chemistry, you can test how much students have been in lab by asking them the color of some chemical. If the student has seen sufficient lab time, he will know. In this way, they can test more than simply book knowledge (the assumption here is that no one actually sits and learns the colors of different chemicals). Fields in which concepts are easy to test benefit more. Fields where the tests only measures knowledge tend to be much less useful.

  4. The point is that everyone learns slightly differently, and that may mean for some a standardized situation is best, while for others a standardized situation is the worst possible circumstance.

    Which is why school vouchers would be a good thing. Falling short of that, at least there should be more charter schools. Given that the teacher's unions hate either ideas, it's going to be a tough battle. In the meantime, expect all sorts of standardization and other ways to make politicians feel good about themselves.

  5. I think macneil got it. People qualified to teach a high school course in computer science could probably make more money elsewhere. Despite the incentives given to teachers (summer vacations, tenure, the ability to work toward a master's degree while teaching, at least in my state, etc.) I think math and science teachers are still going elsewhere. Salaries probably aren't high enough, and this has the dual effect of giving teaching a bad social stigma. I think college-aged people and young professionals view teaching as a way to sort of give up. And if liberal arts and fine arts teachers make as much money as those in mathematics and the sciences, where there are more lucrative opportunities, there is bound to be even less competition for science and math teaching positions. My high school calculus teacher (who was great) became a math teacher after doing much of her undergrad work in English for this very reason. Higher salaries would lure even more competent teachers, teachers who genuinely liked math and science, into teaching. Then, maybe some of that enthusiasm would rub off on students and you could all have more computer science geeks to choose from.

    The AP exams' main problem isn't that they're standard. It's that they're so canned and risk-averse. They heavily emphasize questions that can be impartially graded. This is probably because admissions people tend to prefer data that is bell-shaped and complaints are unavoidable when so many years of students' lives hinge on tests that take a few hours.

    I really liked the European IB standards, and I found that they allowed both teacher and student much more flexibility in the course. Most courses that I knew of involved a large project, sometimes over the course of two years, that was sent off to a committee to be graded. I don't know what kind of mechanism they had to deal with disagreements in grading, and that would have to be dealt with to make a comparable American system economically feasible. Right now, though, I think the College Board and ETS expend so much energy avoiding the risk of lawsuits that their tests just end up sucking. I see two solutions: create a national standard for standardized tests, and let the private sector work out the business model for dealing with grading complaints (subsidizing them if no model seems feasible), or, just have some tax-funded public entity create the standardized tests. Because I don't see much competition for creating standardized tests, I don't think the market model is really doing any good; by definition a standardized test maker would probably be a monopoly. I think the second approach would be more efficient, since none of the money would be going into business owners' pockets.

    Who said anything about vouchers???

  6. Hi Josh. I mention vouchers because they are the anti-thesis of standardized testing. Instead of making education some big monster from the top down, it becomes something much more local, where parents have more control.

    If a school is failing a particular student (while, perhaps, doing well with other students) the parents of that student should have the right to take him/her to someplace that's a better fit. If a school is failing everyone, enough parents will wise up and run that school out of business (public or not) to replace it with something better.

    Charter schools can be better for these reasons: at least when a charter school is bad, it goes away. Imagine what malls would look like if every store that got started stayed in business. There would be so many bad stores around, pushing out the good ones.

    Of course, the economic case for public funding of education is clear. I think the case of public funding for private education is just as clear.

    Some are skeptical of the idea: if you give parents "too much" power, soon you'll see complaints from other people who don't like the way other parents raise their children. It's like the mother who says that, say, commericals shouldn't target young children. Of course, this mother could just turn the TV off. But what she really wants to do is turn the TV off for everyone else as well.

  7. Vouchers? Vouchers? Seriously, vouchers are a way to help rich people get richer. Giving my parents a voucher wouldn't have helped anything. We lived in the poor rural midwest. Unless you were going to let the colleges take my high school vouchers, there wasn't a good school around.

    The real problem is that High School==Babysitting until college.

    So let's not talk about vouchers.

    I never took any AP courses either, the school was too poor (or something like that) to have them. Not that many would take them anyway, there isn't demand for helpful things to get into college when only 10% of each class goes to college, and 50% or more of those students go to Ohio State University, i.e. the big behemoth where no one knows your name.

  8. You first claim that vouchers make rich people richer (perhaps you've never heard of progressive vouchers?), and then you talk about how poor your school was.

    Wouldn't you rather have the option to have bussed in to a school 35 minutes away that would have been better? Even if you never took the option that extra choice would have improved your own school.

  9. I think the proliferation of AP courses (especially in areas in like CS, where most of the instructors are not well qualified) takes away from teaching the mathematics, writing, and other basic skills which many of our entering college students lack nowadays. To paraphrase my usual joke, I think, for example, Richard Feynman did fine without AP exams ;)

    And please help us if the AP tests beome another "high stakes" exam on which schools' funding is based (as one poster suggests). The FCAT exam in Florida has seriously hurt the quality of K-12 public education (as one who sees the output of this system).

  10. Responding to Macneil:

    Right now, if you want to send your child to private school, you pay twice, once to the public school system via taxes (that is, if you pay any taxes at all, most corporations get exemptions and most rich people use loopholes, but regardless), and once to the religious, ahem, private school.

    i think this is fine. vouchers enable rich people to not pay the taxes to the school system, and just pay the private school. so yes, by definition, it will make rich people, the ones who can already afford to pay, pay even less.

    to your second point, the bus to my high school took 1 hour and 15 minutes as it was. Other students took bus rides up to 2 hours. Do you really think bussing to an even farther away school is plausible.

    In case you aren't aware, the size of the territory you have to cover goes up with the square of the radius you expect students to travel to get to school. In other words, no efficient and practical bus system is possible.

    But, the rich kids could still take their limousines, it won't affect them too much.

    You can criticize me all you want, you can even say I'm calling for class warfare. That's fine, I am.

    The public school system needs severe fixing, and it is not going to happen when all the funding is pulled. we could just abandon it entirely, but no one likes to think in those terms.

  11. Unfortunately, as I understand it, private schools do not need to accept vouchers. And they can obviously reject students for reasons that public schools cannot. So some students are unable to find acceptable private schools, and their problem is not solved.

    Moreover, some schools simply don't have the resources to compete effectively. Parents might take vouchers to schools that perform well academically, while public schools that are handling money well and could otherwise drastically improve could be left out in the cold. A little competition is fine, but efficiency, not raw academic performance, should be rewarded. And need should always be considered.

    The simple fact is that the whole market vs. public money debate masks a more fundamental question: who gets the money. In NYS local property taxes fund schools, and I don't know of anyone who pays property tax who would be keen on distributing that tax across the state to other districts; they want it to go to their children. State aid is a pittance in comparison, and isn't guaranteed to go to the neediest districts anyway. On the other hand, I think most of us believe that economic mobility ought to be a reality.

    Those two things don't really go together.

  12. To the poster who equates "going to private school" with "being rich", I'd like to point out that there are many people who choose to send their kids to religious (private) schools and can barely afford it. (Or can't afford it at all, and must beg and/or borrow to make it through...)

    Whether or not vouchers should be used to pay for (part of) a religious education is debatable, but your premise is flawed.

  13. I don't want to rehash every point/counterpoint here. That was already done hundreds of times in the 80s. But, as a detail, under a progressive voucher system the "rich" would get less of a voucher amount (if any at all) than the "poor". As for the rest of the arguments, simply substitute "college" with "school" and you'll see how silly some of the claims sound.

    I also think local funding of education systems seems to be best. For one, you want parents to have autonomy and be in control. Creating huge school empires might be fine with administrators, but the larger a school gets the less say any of the parents have. Second, what happens in situations where local funding isn't a major part of total funding is that the richer areas will increase restrictions on development. No new houses will be made in the area and property values will be kept high. The incentives change when things are more locally funded: towns are more receptive to further development if they can receive extra tax dollars from it.

  14. Why had this degenerated into an argument about vouchers? I thought the question here is about the CS AP.

    That being said, I believe the initiative is up to the colleges to tells the high schools how bad the exam is if they think so. Not considering the results is certainly a good start.

    And while we are there, can we fix the CS GRE too?

  15. Some of you have asked "Why talk about vouchers?". The reason is that vouchers are one of the best solutions we can use to improve our schools.

    Some cynically think it has something to do with transferring money to rich people or the state establishing religion. But at heart the only reason economists like the idea is because it makes schools better. All of this standardization business is more a solution that is grasping at straws.

  16. Vouchers allow parents to choose what it is they want their kids to hear. This can be good or bad.

    * if you want ID instead of evolution come to our school

    * if want religion come to our school

    * if you want a more challenging language program come to our school.

    * if you don't want too much homework come to our school

    * if you want bad things said about gays then come ot our school

    The question is: do you believe that parents will make good choices? Given all the surveys about how much American parents know about math, science, and geography, etc I would not be counting on too many good choices.

  17. But your list covers everything any parent with the means to can do *today*. And even very poor parents can choose to homeschool.

    Bill Gates can hire tenured professors to tutor his children. They'll get an amazing educational experience that no one else can have. Should everyone get what Bill Gates's kids get? No, that's just not practical. Is that fair? Of course not. It's also not fair that Brad Pitt is better looking than average.

    Should Bill Gates's kids be forced to attend public school? That sounds like a horrible idea to me. That's like forcing Brad Pitt into plastic surgery to make him look average.

    Anyway, most good things have their costs. Fire sure is useful, but it's dangerous. I don't think parents making bad decisions in terms of school choice is going to make the children any worse off. However, not letting parents choose will do much more damage... the damage of lost opportunities that are both very real and something people don't notice.

    I also have a lot of faith in the capabilities of parents to parent well. (As for the really rotten ones? Child services already is in place to take care of that.)

  18. It's slightly off-topic, but see this rant from Joel Spolsky, The Perils of JavaSchools.