How early for APCS

In what grade should students take APCS? This question comes up from time to time.

I've heard answers ranging from middle school through never. Infact, years ago, my chairman relayed a conversation he had with Marvin Minsky where he asked Minsky what the high schools should be teaching with respect to CS. The answer was "nothing." This was then amended to "teach them to type." Of course this was a long time ago but I believe the sentiment was that college was the right time and the high schools don't know what they're doing and will just screw up the kids. Actually, I still see some of this attitude today and it gets passed down where high school teachers sometimes don't want the middle schools to "mess up their kids" and on down the line.

At this point, it's pretty clear that you can do good CS at the high school level and as things are being pushed down the grades we'll eventually figure out what's right and when.

Still, within the high school the question of what grade for APCS-A remains.

There isn't a single right answer for this but I can share my experience and what I ended up deciding when I developed things at Stuy.

As parents, my wife and I learned early on that we'd have to supplement what our kids would learn in school, particularly in terms of math and science. I was in charge of math. Since I used assorted books to help, math time came to be known simply as "MATH BOOKS." It's funny that now while my kids don't look back super fondly on MATH BOOKS they agree that it was an important thing for us to do.

I discovered, as I'm sure many parents do, that there are times when a kid is just not ready for a subject. At one point I tried to introduce Algebra too early and saw it wasn't going to take so I pulled back. A while later, we tried again. This time it was clear that they could do the mechanics and solve problems based on rules and formulae but they really didn't "get it." A third attempt some time later, they were ready. They mastered the subject with deep understanding easily. It could be argued that the early exposure helped but I don't think it did.

The takeaway here is that kids can do the mechanics and appear to succeed but if it's too early, they don't really learn the subject.

I saw this on a larger scale with "Physics First." At Stuy, most freshmen would take Bio in 9th grade followed by Chem and then Physics in their Junior year. Physics first kids would take Physics in the 9th grade. I don't have any evidence for this but based on my inquiries over the years it seemed that the kids who took physics in their junior year had a deeper understanding of the subject.

I also personally saw this when teaching math classes early on in my career. Most Stuy freshmen take geometry. The math aces take precalc or calc but a few weren't math aces but were the star in their middle school. They were pushed through geometry in 8th grade and started Stuy in Algebra 2 and Trig. The majority of these kids that I taught could spit back all the formulas and theorems from geometry but very few were actually ready for Alg2/Trig and struggled considerably.

So what about APCS? At Stuy we always taught a superset of the old APCS-AB. It's mostly APCS-A in the Fall semester and Data Structures in the Spring. When I started, there was no prerequisite for APCS so I had students in all grades. Mostly Juniors and Seniors but a few Sophomores. I think I only had a freshman in APCS once and they were an outlier.

Some of my sophomores were ready, more could merely spit back material and do the mechanics - enough to pass the APCS-AB exam but they really didn't get it. There were some, however that weren't ready. These were very bright kids and while I can't say for sure, deep down I very much suspect that if they took my class a year later they would have done much better and they might have followed a very different path towards their future.

As a teachers we can change lives for the better and in fact save lives. We can also do great damage. Because of the latter, teachers should follow the mantra "first do no harm." This led me to a general rule of taking kids into APCS-AB until their junior year. If I were to allow a sophomore in I wanted to be pretty sure that the kid was ready.

Later we were able to create a sophomore year CS requirement which gave all kids exposure to CS prior to APCS-A (we still taught AB but by now only the A exam was offered) and also gave us a platform to vet the outliers who might be ready for APCS. For those kids, in addition to an interview we had them self study APCS-A, do a project and sit for the APCS-A final exam and then we'd add them to the second half of our AP class.

By and large the system has worked. Kids now get an intro in the 10th grade followed by APCS in 11 and more electives if desired in the 12th. For the kid that's truly ready for APCS in the 10th grade, we have a path. We don't make it easy but the path is there.

Over the years we've had some CS superstars come through the program. Some have pushed back – "why do I have to take the intro?" "Why can't i start right away in APCS?" – after all was said and done, all but one came back to say that we did it the right way and that they benefited from all our classes.

At Stuy, it seems that for most students, 11th grade is right for our APCS. Between academic maturity, other classes, scheduling and everything else, it seems that this works the best for us.

Frequently the question as to when to teach APCS revolves around how early it can be taught. I have to ask - why are we so caught up with doing things so early?

If we're just pushing college level courses down to high school then what's the point? The kids will have to take the classes again if they're going to major in CS anyway.

If you have a program that doesn't duplicate college courses you might want to have kids finish APCS-A earlier but I wonder how early you really need.

As a society we keep pushing things down younger and younger. We're pushing AP classes like history or APCS-P down to 10th grade or even earlier. You have to ask, is a course that's developmentally appropriate for a 18-22 year old developmentally appropriate for a 14 or 15 year old? Probably not. You also have to ask that if it makes more sense to push further ahead or if it's better to do enrichment at a level the kids can handle.

I'm an enrichment over acceleration guy but other people differ. Just like the question of when for APCS, there's probably not a single right answer to this.

As I said up front, there's no right answer to this. I'm still convinced that 11th grade is probably the best general entry point for APCS-AB but if I were only teaching APCS-A maybe I'd shift it down a year. My gut tells me no, partially based on teaching non APCS to 8th, 9th, and 10th graders but maybe. Of course, we all teach different students in different environments so your mileage may vary.

There are also those outlier kids that can truly master the subject at an earlier age, I'm not really considering them here but they should be accommodated when possible.

If you've been running a program for a while, look at the data - not how many kids pass the exam - look at how many master the next level material. If your kids next class is data structures in college, find out how well prepared they were and see if and how it maps to the grade in which they took APCS. If you're just starting, I'd recommend erring on the side of caution - remember "do no harm."

If you have to teach CS and it turns out you have to teach it when the kids are too young to master APCS-A then don't teach APCS-A, teach great CS at a level that's appropriate.

All of this will work itself out in time. Until we get there, there will be some mis-steps - I know I've had mine. As long as we continue to move the kids forward and give them something that they wouldn't have been getting otherwise while we figure all this out, we're doing pretty good.