Comments on: Collaborative Floundering trumps Scaffolding

By: Khan Academy offers kind-of-scaffolded computer science learning: Doing away with the teacher « Computing Education Blog

Wed, 15 Aug 2012 12:33:03 +0000

[...] don’t see it as a revolution in computer science education — not yet, anyway. Now, maybe it’s way of supporting “collaborative floundering” which has been suggested …. Maybe they’re right, and this will be the hook to get thousands of adolescents interested [...]

By: Shuchi Grover

Shuchi Grover — Wed, 02 May 2012 23:48:15 +0000

Indeed, Stephen. I presented the results of the first couple of studies at AERA a couple weeks ago, but they don’t publish the proceedings, unfortunately. I am in the process of building out my webpage on the Stanford server where I hope to upload the paper. I am working with Dan Schwartz on this, and I hope to run the next iteration to investigate the mechanisms at play (self efficacy or learned helplessness or …?) and also test this in other problem-solving contexts besides anagrams. Hopefully then it will be publication-worthy

By: Shuchi Grover

Shuchi Grover — Wed, 02 May 2012 23:42:01 +0000

Yes, we figured that these findings would be especially useful in settings like games and intelligent tutoring systems.

By: Stephen Downes

Stephen Downes — Wed, 02 May 2012 18:44:20 +0000

Shuchi, it would be a great idea to post your findings online – I searched for them and found nothing – unless they are online they are basically invisible.

By: Mark Guzdial

Mark Guzdial — Wed, 02 May 2012 12:54:57 +0000

I wonder if it's like the worked examples effect: Worked examples are great for low-knowledge learners, but can hinder learning as students advance.

By: Aaron Lanterman

Aaron Lanterman — Tue, 01 May 2012 19:14:29 +0000

This sounds analogous to the sort of issues game designers struggle with. What’s the best way to ramp of the difficulty curve?When (and how) should the game give “hints?”

By: Bijan Parsia

Bijan Parsia — Tue, 01 May 2012 19:00:37 +0000

“Floundering” seems like exactly the wrong term (“productive failure” seems quite good).

And indeed, the article is way more circumspect:

This has led some researchers to argue that instruction should be heavily guided, especially at the start, for without it, learning may not take place (e.g., Kirschner, Sweller, & Clark, 2006). Further support for starting with greater structure during instruction with a gradual reduction (or fading) over time as learners gain expertise comes from several quarters (e.g., Hmelo-Silver et al., 2007; Puntambekar & Hübscher, 2005; Wood et al., 1976)…

What is perhaps more problematic is that an emphasis on achieving performance success has led in turn to a commonly held belief that there is little efﬁcacy in novices solving problems without the provision of support structures initially. In contrast, our work is grounded in the belief that engaging novices to try, and even fail, at tasks that are beyond their skills and abilities can, under certain conditions, be productive for developing deeper understandings.

Looks like a great article! I’m getting some popcorn…

By: Christine Alvarado

Christine Alvarado — Tue, 01 May 2012 17:32:25 +0000

I have seen similar results (from Math education) in the past, and it does make sense to me. But one thing I always wonder is, learning outcomes aside, how do students feel about the subject matter when they are done “floundering” (or being scaffolded)? In my annecdotal experience (and others’), students really don’t like to struggle. I worry that they might get turned off to the whole field if the struggle is not very carefully designed and controlled, even if they are learning more. Given that so many students don’t like CS in the first place, this could be a very bad thing. Are there any studies that have looked at persistence over time with these different approaches, rather than just learning gains?

By: Jeff Rick

Jeff Rick — Tue, 01 May 2012 16:22:07 +0000

I had a chance to grill Manu on productive failure for an hour and “floundering” is not an accurate portrayal of the concept. I recommend reading the original article. Basically, the problems he assigns are very well designed so that learners truly engage the concepts when they attempt the problems. They think they know how to solve it, but then fail because the problem is more difficult than it first appears. They experience expectation failure (something Janet would say is tremendously important to learning) and start to see what solutions don’t work and why not. Probably your floundering Java students do not experience expectation failure and are unable to use their experience to deduce what does not work.

By: Shuchi Grover

Shuchi Grover — Tue, 01 May 2012 15:12:13 +0000

This is very interesting, Mark, and a coincidence too! This past AERA I presented findings from a set of studies titled “The Hidden Cost of Help on Subsequent Task Persistence”. (The idea of perseverance came from this belief – and observation – that it is one of the hallmark dispositions of good programmers). The task involved anagrams, to be solved individually – with or without hints (of different kinds).. and persistence was measured on a very tough transfer task involving an (almost unsolvable) anagrarn.

Going back to the article, though, if the “floundering” was not well designed, then the results may well have been the same as that of your students programming in Java. “Designing for productive failure” may not be so trivial, and while I can see the benefits of “hidden efficacy” and such, I’m not sure of the findings can be applied very easily to every learning context. And on the flip side, “scaffolds” too come in various shapes and sizes – some can impede learning whereas others I’ll bet can work as well or maybe even better than “designed floundering”. In my study I gave “direct” hints and “strategy” hints. The direct hint group quite obviously did the worst, whereas the generic strategy hint group did marginally better than the group that received no hints at all…