The ACM ‘paywall,’ computing education research, and open access

I reference research papers regularly in this blog, often in the ACM Digital Library. I’ve been receiving more complaints lately when I reference papers “behind a paywall.” After I linked to the article that Leo Porter, Beth Simon, Charlie McDowell, and I wrote about successful practices in CS1, someone tweeted that we were “whores” by allowing our paper to be sold by ACM. As Greg Wilson said to me, the support for open access in our community is “vehement.” Now, there is a petition demanding that the ACM open up the Digital Library, free of charge.

I’m a computing education researcher in the ACM SIGCSE community. “Open access” is much more complicated in my community. The arguments for opening access are more subtle in under-funded and even non-funded education community.  The British Academy has just released a set of papers (July 2013) on the challenges of fitting social science and humanities research into open access models.  They argue that we need a ‘mixed economy’ because there are different expectations and funding models for research in different disciplines.  Open access is different for computing education research than other areas of computer science because it is a social science.

Why Education is more complicated for Open Access

The case for open access is made in the first sentence of the petition:

Computer science research is largely funded by the public, for the public good.

There are two cases to consider: the research that is funded by the public, and the research that is not. Let’s start with the research that is not funded publicly, because that’s a big part of what makes education unusual.

Many (maybe most) of the papers published at the SIGCSE Symposium and the ICER Conference are not supported by public funds. Go through the SIGCSE papers and note which reference public funding and which don’t — it’s a pretty high percentage that don’t. ICER was created explicitly because there were groups of faculty, without public funding, who were collaborating and doing experiments in their classes and then pooling the data. They needed someplace to publish. Those faculty were paid to teach, and they had heavy teaching loads. They did the research on their own time, because they valued doing it. I don’t see how the public can lay claim to their work.

Some of the work at SIGCSE is publicly funded, but maybe at lower levels compared to funding from Department of Defense, National Institutes of Health, or Department of Energy. My research is funded by the National Science Foundation (NSF). How much we are funded is limited by NSF rules and by institute rules. For example, graduate student research assistants can only be paid for up to 20 hours of work (and only 15 before passing qualifying examinations, in my school). Few PhD students complete his or her research work in only 20 hours a week. Let’s say it’s 60 hours per week. Are we really arguing that all of that student’s work is “funded by the public” when that is true for only 1/3 of the hours? Should the public be able to lay claim to all of the student’s work because of those 20 hours per week? If anyone does work outside of what they’re paid for, isn’t that their work?

The issues are actually much the same for faculty, though we get paid much better. Faculty at my school are funded for 9 months by the state of Georgia, and I do federal NSF-funded work for an additional two months per year.  For the last two years of “Georgia Computes!” I could only charge two weeks (specifically, 80 hours) of my time to that project per year. ECEP is a five year project on which I can only charge 160 hours per year. I spend 150 hours per year just on the management meetings for ECEP. I’ve already spent more than 40 hours on the road, doing the work of ECEP in Maryland, South Carolina, and at the CSTA Conference. All of that is before the work of evaluating data and writing papers. I am pretty sure that the state of Georgia does not see itself funding my work with these other states. Simply put: The federal government does not fund everything I do. If they don’t fund everything I do, I don’t believe that they can lay claim to it.

When I shared this story with my colleague, Beki Grinter, she pointed out that the case is similarly murky for corporate-funded work. Microsoft paid for the robotics CS1 work here at Georgia Tech and Bryn Mawr. Can the public lay claim to that work, too? That work is in the ACM Digital Library. By what right is that work made freely available?

All authors want their work to be distributed widely, to have impact. I usually provide copies of my papers when asked, and I use the ACM DL Authorizer service to provide free access to my papers. It’s up to other authors to decide if they are willing to do the same. Yes, opening up the DL would allow the papers to be distributed even more widely. But is that sustainable? What about the funds that are lost? I am willing to forego that breadth of access in favor of the good of closed access. That’s the deal that all the ACM authors made when they assign ACM copyright. The open access movement aims to change the agreement, after the fact.

Education research and the developed world

The petition I mentioned earlier focuses on the public funding for the public good. Another argument for open access that I’ve heard (and thanks to my college, Ellen Zegura, for helping me understand this) is to serve people the developing world — people who don’t have access to the resources of the developed world, and for whom ACM Digital Library access is prohibitively expensive.

Education research is different than most CS research because it’s a social science.  Are the papers published in the SIGCSE Symposium and the ICER conference directly useful to the developing world?

• Education is a process of acculturation. It’s about passing on culture from one generation to another.  If you change cultures, you have to change education.  Please take a look again at the The Chronicle article I referenced in this blog post about designing learner experiences.

“Transferring education from the United States to Africa wouldn’t work,” argued Bakary Diallo, rector of African Virtual University. “Because we have our own realities,” he added, “our own context and culture.”

• Writing humanities and social science research is a dialogue with an audience (as described in this piece in The Guardian).  It’s not merely a process of reporting findings.  If you are writing for a developed world audience, you are explicitly not speaking to a developing world audience.  If you want to write for a developing world audience, you should learn to write for that audience.

I have not worked in the developing world, so I can’t speak to the issues of bridging the gap between the developing and developed worlds.  But most education researchers have faced these issues of differing cultures and audiences.  I have talked about Media Computation in several countries.  When the places I visited were like my culture and audience, it worked pretty well — MediaComp is being adopted successfully in Australia, for example.  When the places I visited were not like my culture, I realized that I was solving completely the wrong problems for them and what I was saying was useless.  When I spoke to teachers in China and Mexico and Qatar, I realized that I needed to listen before I could say anything worthwhile to them.

The problem of transferring education research isn’t just a problem of the gap between the developed and developing world.  In ECEP, we are realizing that even curricula, outreach programs, and policy approaches don’t transfer between states — even neighboring states!  I work in Georgia, South Carolina, Massachusetts, and California now.  The values and concerns are very different even between Georgia and South Carolina, and we’re really struggling to figure out what our summer camp model means in Massachusetts and California.

There’s a perspective that says that this view is “patronizing,” and continuing an “us/them” perspective. I believe in tailoring for different audiences, but that doesn’t imply superiority of one audience over another audience.  The key idea in my work is that one size does not fit all for computing education. In our CS classes, we often make the mistake of assuming that what works for some percentage of our class is good enough for everyone, and if some don’t succeed with that approach, it’s their fault. There is evidence to believe that different students succeed best at different approaches, e.g., that there are aptitude-treatment interactions,. Cognitive science has told us for decades that students’ prior background influences how and what they learn. Our Media Computation approach improved the success rates of liberal arts students at Georgia Tech, from a less than 50% success rate to an 85% success rate.   I don’t believe that my liberal arts students are superior to my CS students, or vice-versa, but I do believe that each group has different goals and succeeds best with different approaches.  I’m concerned that pushing for open access is making the same mistake that we keep making in CS — if it works for us, it’s good enough for them, so just give it to them and let them figure it out.  (Kind of like MOOCs.)

Any responsibility that the developed world has to share research with the developing world is not met by simply sending them our papers.  If we want to share our research findings, we have to learn their educational problems and their educational goals and values.  We would have to learn to communicate about their issues.

Where does the money go

I have to admit a bias here: I consider myself part of the ACM community. I value being part of that community, being an editor and reviewer and author, and that funds from those efforts goes to sustain the community. Language matters — ‘paywall’ sounds permanent, as a “wall” is. It’s really more like a ‘tollgate,’ where the tolls support the community.

The ACM does good with the funding it receives, from my perspective in education. The funds generated by the DL go back to support the authors’ research communities

• A portion of all fees generated from SIGCSE publications goes back to the SIGCSE Board. I have served on that Board for the last three years. The funds are used for travel grants to new faculty to get them to the SIGCSE Symposium, for special projects funding to produce new curricular materials for the community, and to provide for a rainy-day fund in case conferences don’t break-even. If the DL funding wasn’t there, the conferences would probably have to raise their rates, to reduce the risk of ending up with a deficit.
• ACM itself funds efforts like the ACM Education Board and Education Council. These organizations fund the development of curriculum standards. By “fund,” I mean pay for travel, food, and lodging. The participants volunteer hundreds of hours of their own time for a really important purpose. These curricular standards are particularly important in the developing world, to serve as a guide for what a CS degree is supposed to be.

Of course, part of the fee goes to maintaining the DL, and that’s not insignificant. I hope the DL will continue. That costs money. A fee-based system is sustainable.

The ACM is not a nameless corporate entity. It’s a volunteer-driven, membership community. The DL is not a bank that is covetously hoarding intellectual wealth. It’s a source of knowledge for computing professionals, and a source of funding for the good work of ACM.  If we want to make our research findings useful elsewhere, we should actively do that by understanding those cultures and audiences. We cannot expect that creating open access will necessarily fix educational problems elsewhere, but demanding open access may cost our community a lot.

(Thanks to Ian Bogost, Briana Morrison, and Leo Porter for advice on an earlier draft of this post.)