Mark Guzdial's group at Georgia Tech does world-class work on how to teach programming: over half of the links in the recommended reading for our instructor training course point at Mark's blog, and I've learned much of what I know about education from following up the pointers he has provided.
Two and a half weeks ago, he posted a link to a paper that he and others had written summarizing what actually works in introductory programming classes. The paper itself was behind the ACM's paywall, and reaction to that ranged from disappointment to vulgar.
Mark responded thoughtfully, arguing that the case for open access in the humanities and social sciences (including education research) isn't as clear as it is in the physical sciences and computer science, where we have now reached a tipping point. In particular, he pointed out that most computing education research isn't publicly funded, so the "pro bonum publicum" argument doesn't apply.
In my mind, though, the real argument for openness is this: if you want your research to have an impact outside academia, it has to be openly accessible. Mark himself made this point just a week later in a post that quoted Jai Ranganathan:
...if you and your colleagues don't convincingly make the case to the public that your discipline should be funded, well then it won't be.
Putting something behind a paywall is exactly the opposite of "making the case to the public". Of the eighty people working in Mozilla's Toronto lab, none have legitimate access to paywalled research, and as far as I could find out, none had read a paper from a closed-access academic journal in the last twelve months. Experience has taught them that if clicking on a link brings them to a paywall, it's more fruitful to hit the "Back" button and try the next link in their search results than to try to find a way around or through the paywall. An economist could argue that they ought to pay $20 a paper on the chance it might be useful, because if even one in fifty saves them a month of work, they'll be ahead, but most people don't think that way.
In economics, Gresham's Law states that bad money drives out good. The Internet corollary might be dubbed Wales' Law: open knowledge drives out closed, even when the open knowledge is of lower quality. Given a choice between a freely-accessible Wikipedia entry or Stack Overflow post, or a peer-reviewed report of a careful multi-year research project that costs time or money to get to, the overwhelming majority of people are choosing the former.
But nothing's really free. The true price of closed access is people making poor decisions: many of the entrepreneurs, programmers, and politicians involved in "education and technology" ventures don't even know what they don't know about education. Open access alone won't fix this, but putting research behind paywalls certainly doesn't help.
Closed access to research is a relic of an earlier age in which information was moved by atoms instead of by electrons. Like all of us, publishers and professional societies are struggling to come to terms with a rapidly-changing landscape. I agree that it's important to ensure the vitality of the ACM and other professional bodies, but I think it's equally important for research to have real-world impact. And that, I think, is the strongest argument of all for open access.
Postscript: Alex Rudnick independently made a similar response yesterday on his blog.
Further postscript: The September 2013 issue of IEEE Computer is devoted to computing education. If I've done my math right, it'll cost me $95 to read the five articles it contains on the topic. In the immortal words of Reverend Horton Heat, it ain't gonna happen.