Thursday, November 22, 2018

The Problem With CS Education Research

There is some research in who t teach computer science (OK, mostly its about teaching programming but that’s important too) but not a lot of teachers or professors are adopting it. One has to wonder why? Eugene Wallingford takes a look at this question on his blog Why Don't More CS Professors Use the Results of CS Education Research?  He has so reasonable answer having to do with peer pressure and pressure from students and parents to do things “the way we’ve always done it.”

The path of least resistance is tempting of course. And change is hard. But I think there is more to it than that. I mentioned Parsons Problems to a professor at a university and he knew nothing about it. And why would he? For many professors at R1 schools research is a higher priority than teachings. And teaching high level courses has a higher priority than lower level courses.  But higher education is not my area so I look closer to home – secondary school teachers.

I should start with myself though be fore I look at others. Problem number one for me is lack of understanding. I’ve read some papers on worked examples and Parsons Problems. I think I understand them. Sort of. But the language of research papers is not one I am totally comfortable with. No, that’s not quite right. I read it comfortably but figuring out how to put the conclusions into practice is often still a little opaque to me. These papers are written for other researchers and not for people to implement in their own classrooms. So that is problem one.

Then there are the tools that are used in the research. I have played around with a couple of Parsons Problems tools. Eh. Clearly they were developed by people who were developing for a specific audience that is possibly more forgiving or willing to work harder than the average high school student.   Lots of them look like the old school (i.e.. before windows) applications adopted for the web. Hardly the thing to grab the attention of a modern HS student. To be honest I tend to worry about using tools developed for research projects. One never knows when the researcher will move on to a new interest.

I want to use Parsons Problems but I struggle to find the right tool to use. I want to use worked examples but I need to spend sometime with someone (or some paper) that knows how to help a HS teacher learn best practices. I’m sure there are other things I can and should learn from CS Ed researchers but I don’t know what they are.

That brings up the next problem. I spend money out of pocket to be a member of ACM and SIGCSE. That gets me a lot of great papers to read. Money well spent IMHO. But it’s been a while since I have been to the SIGCDE conference (maybe this winter) and that would let me get more out of these papers. I could ask more questions and hear the questions of others. I think a lot of HS CS teachers don’t really even know much about SIGCSE or attend the conference. Yes, there are a bunch of HS CS teachers at SIGCSE. They’re a small subset of the total community though.

There are more HS CS teachers at the CSTA Conference but that’s not a venue high on the radar of university CS Ed researchers. There are other conferences more likely to help a professor get tenure or impress possible funding agencies. That is not to suggest that CS Ed researchers don't care about HS CS. They absolutely do. Several have been very generous in their support of CSTA and of spending time helping HS CS teachers. I owe them a debt.

Many HS CS educators don’t have the time, energy, or motivation to go digging for new ways to teach. It’s hard to blame them. Their students do well on examples, pass the AP tests, do well in university, and generally seem happy with the statues quo. How are they ever going to learn how to do things better or even decide if there are better ways?

In short, yes, some are unwilling to try new things but more would if a) they had a better understanding of the techniques and b) they had tools more suited to younger learners. The biggest factor though is they just don’t know about the research and what it suggests. Someone needs to take that on. Calling NSF, Code.ORG, CSforAll, SIGCSE, and CSTA!


Michael Ball said...

I'm not going to dispute that research to practice is difficult. Of course it is, for the variety of well known reasons. I'm not 100% sure that CS is worse than other subject areas, though I'd certainly be interested to see data.

However, I do think we have to look more broadly at "research". Look at Carol Dweck's work. She started working on it around 2003 IIRC), or thereabouts. It took nearly 10 years for it to get big and influence courses, but now many people know about it. In part, it's because her research has grown so popular, but I knew people who were talking about it even when I was just getting started.

There's also the fact that all the new CS teachers being trained are mostly using curricula and PD based on recent research. Whether it's BJC or Bootstrap or -- we're all using and contributing research that gets put into practice. Certainly, we could do better about highlighting what's based on research or teaching teachers how to adapt their courses, but they're already plenty busy and we don't have as much time as we want.

By and large, academic research moves somewhat slowly, especially since academics tend not to really believe something until it's been replicated or cited many times over. (Look at the comments in that thread about poor paper quality as one example.) I think it's reasonable to expect a long timeline for research to turn into practice.

Your right that the parsons problems aren't easy to implement today. But, honestly, we all know it's not too difficult to make them easier to implement in the classroom. It's that there's never been (enough of) a reason for researchers to spend the time.

We need to do better as a community, but I think it helps to recognize what research we do put into practice first. Then we need to figure out how to change the incentives -- research HED research coming from (primarily) R1's, we're going to be battling the same issues as any other field, but I think there's some things we can do. Why doesn't SIGCSE have an award for research that's easy to adopt in a class / ready to go? Putting up a paper? Then get attention if you include good supplementary materials!

Garth said...

The National Council Teachers of Mathematics (NCTM) has a monthly publication full of practical teaching ideas for the classroom. Ideas that can be implemented with little work required by the teacher. The ideas are submitted by math teachers so they are pretty much tested in the classroom before they hit the magazine. CS needs some kind thing like this. Most math teachers are members of NCTM. Most CS teaches are not members of anything. The CSTA simply does not have the following or the resources for the classroom teacher. Too bad. It probably has to do with the numbers, there are a lot of math teachers.

A lot of the research I read is just that, research. It is not implementable ideas that are ready to go for the classroom. Very little "here, try this and see how it works". For most CS teachers CS is a minor part of their course load and the time required to implement research "ideas" is not there (or even read research). Most of the CS teachers I know are still in the "cookbook" mode, they follow a canned curriculum with little variation. They simply are not knowledgeable enough to stray from the curriculum. A publication like the NCTM Mathematics Teacher for CS might temp many of these novice CS teachers to try something out of their limited comfort level.

rwhite5279 said...

This is a great topic, and one that is always worth of exploring. I enjoyed reading Wallingford's post, but had to laugh at this line: "Some disciplines, such as physics, seem to have developed more of a culture for incorporating education research into the classroom than computer science." As an AP Physics teacher, I know very well that I and my colleagues are still struggling to incorporate research-informed pedagogical improvements into our high-school teaching. I've even had one former student go on to perform his own university-level research on improving physics education, inspired (perhaps) by experiences in my classroom. I've got my "mea culpa"s prepared.

The subject of Computer Science / Computer Programming, for the most part, doesn't even have a well-defined curriculum yet, let alone a coherent set of pedagogical strategies that we can implement with confidence. (This has been my experience anyway.) I agree with Alfred's call for better communication, although I have to admit there may be resources out there that I haven't yet taken advantage of. Time, as has been noted, is always an issue.

Regarding "Parson's Puzzles," I was intrigued by the idea and ended up implementing a paper-based activity in my Intro to CS course, documented at . If any readers are interested in adapting it to their own classroom needs and giving it a try, I'd be interested to of their experiences.