Wednesday, August 31, 2016

OpenStax: Doing OER effectively and sustainably.

I recently discovered OpenStax and their collection of open source textbooks for classes at the high school and collegiate levels.  Their texts hit the major courses in math, science, and the social sciences—from calculus to physics to economics.  For readers of this blog, largely interested in mathematics, the three-semester calculus text, derived from Gil Strang's Calculus, is a high quality text rendered in both pdf and html (and soon in print).  Having taught calculus off and on for the better part of 25 years, this free text is at or near the quality of Stewart's.

Having also contributed a bit to Guichard's Calculus book, another good open source calculus text, I was particularly impressed by the high-level of production apparent in the OpenStax calculus book. I know it's pretty difficult for a volunteer corps to create such a high-quality text.  Curious about how such a high-quality open source text came to be, I reached out to Anthony Palmiotto, Editorial Director at OpenStax, to learn a bit more about how they do things there.

The OpenStax organization is a non-profit hosted at Rice University.  Its first open source text was released in 2012.  Since then their library has grown to around two dozen texts.  The consistency of presentation and quality across all of the texts reflects a high quality editorial process.  Their editorial philosophy is characterized by a focus on larger market classes like calculus, intro physics, intro biology, economics etc.  These are the areas where accessible, high-quality texts can have the most impact.  They focus on a single offering in each of the course areas. They want their textbooks to match closely the offerings in the traditional commercial markets for these course texts. Doing so makes it easier for textbook adopters to consider open source texts right next to their commercially offered competitors.

To begin, each text has primary authors and contributing authors.  The authors and editors are paid for their work (unlike in many open-source publishing efforts).  There is a development phase and a maintenance phase.  During the maintenance phase, errata are collected from readers, authors and editors continuously.  Corrections are made incrementally to the on-line version and full updates to the texts (pdf) occur between the traditional academic terms (semesters) of most colleges and universities.

Baseline funding comes from Rice University.  Additional funding for specific book projects is sought on a case-by-case basis through grants and donations.  OpenStax has received funding from a number of philanthropic organizations including the Bill and Melinda Gates Foundation.  They also receive funding from companies in the educational technology sector.  For example, WebAssign bundles OpenStax texts with its commercial offerings and supports OpenStax in exchange. OpenStax does not make money on the print book sales.  For the sake of accessibility, print books are sold at or near cost.

Though it is notoriously difficult to assess uptake and adoption of open-source texts, the data that OpenStax receives, from its commercial partners, as well as from self-identified users, shows significant year over year usage growth according to Palmiotto.

It should be noted too, that in the case of OpenStax, I do not use the term open source lightly.  In the true spirit of borrowing, sharing, modifying and redistributing, the source for the texts is hosted on the OpenStax CNX platform.  Users are free to fork an existing text and to modify it to suit their needs.  While OpenStax will conveniently host your forked text on their site, you are also free to download, modify, and redistribute the text on your platform of choice.

In the end, though it was hard to get an accurate figure, the total cost of development for the OpenStax calculus text was between $500,000 and $1,000,000.  This goes a long way towards satisfying my curiosity about the high quality of the text.  In other words, it wasn't a volunteer effort, but rather a paid group of authors and developers.  It remains an open question whether a purely volunteer corps can produce a text of the same quality.  However, unlike a typical commercial text, the OpenStax calculus text could represent the foundation of a purely volunteer effort at maintenance and continued development moving forward. Regardless, all of the texts are significant contributions to the OER effort and help to lower costs and increase accessibility for students taking these courses.

The OpenStax approach to open-source academic publishing is effective and sustainable.  It is a model for how open-source publishing can generate high-quality texts that can compete with commercial offerings.  In a subsequent posting, I'll discuss OpenStax's partner program for educational institutions that wish to make open educational resources a priority.

Thursday, August 4, 2016

Expii creator hopes to harness the power of the OER community.

Po-Shen Loh at Carnegie-Mellon is perhaps best known for his success coaching the US Math Olympiad team.  However, he is equally passionate about a project he started called Expii that he hopes will provide self-directed learning of mathematics using open-source contributions and a sophisticated dynamic learning system.

Mathematical subject matter is organized in a collection of nodes that are determined by a core group of editors at Expii.  Navigating through the nodes eventually leads to a specific topic node.  Within that topic node resides a collection of explanations and exercises for that specific topic written by a volunteer contributor.  Explanations are constructed on the site's easy to use editor which recognizes LaTeX.  The site also allows one to embed instructional YouTube videos.  Exercises too are easily constructed within the site's builtin editor.

There are two aspects to Expii that separate it from other open source on-line learning systems, the editorial approach and the dynamic learning algorithm.

First, unlike the Wikipedia editorial approach, contributors do not typically edit other contributors' work.  Factual corrections to an existing contribution are welcome, but stylistic re-writes are discouraged.  Rather, if a contributor does not like the style of a particular explanation, that contributor is encouraged to submit a competing explanation in the style that they prefer.  Subsequently, the two (or more) explanations then compete with each other by receiving upvotes/downvotes from users of the site.  Explanations with more votes are listed first when a learner visits a particular node.  Prof. Loh hopes that this approach will provided a variety of explanations on different topics that will resonate with different learners.

Second, the learning algorithm is a novel application of the ELO ranking system that is used in international chess competition.  In a sense, learners and exercises are in competition with each other and are ranked using an ELO-type system.  In this way, when a learner requests an exercise on a particular topic, their ranking is matched to an exercise with a nearby ranking.  As the learner improves, their ranking improves and they are presented with increasingly challenging exercises.

At the moment, the Expii effort is funded by private investors and is a for-profit entity.  However, Prof. Loh says that the future business model will not include a paywall for access to the site.  Rather, one possible business model would be mediating the pairing of contributors and learners on Expii into paid, adhoc, teaching and tutoring relationships.  The core group is looking for investors and applying for federal grant money to keep the 3-year effort going.  There has been steady growth in both users and contributors, but it seems to not have reached "critical mass" yet.

The site welcomes the addition of existing open-source materials as long as they are appropriately licensed.  For example, a Creative Commons license that allows commercial use would work.

Prof. Loh states that a lot of his motivation for starting the site came from his work with the Math Olympiad and other mathematics competitions.  He recognizes that for the US (and other countries) to compete, there needs to be a large pool of talented mathematics students.  He hopes that Expii can deepen that pool.  Indeed, he hopes that it can take a struggling mathematics student that may not be getting quite what he/she needs in their home classroom and provide just the right kind of learning experience to turn them into future Math Olympians.  And, as a great side benefit, Expii will elevate the mathematical literacy across the board.

If you have OERs that you have developed and would like a venue, Expii may be a good choice for you.  The ranking system may be the kind of objective assessment of your work that could be used in tenure, promotion and merit considerations at your institution.