Librarians and the Era of the MOOC

The New York Times dubbed 2012 “The Year of the MOOC,” and it has since become one of the hottest topics in education. Time magazine said that free MOOCs open the door to the “Ivy League for the Masses.” Two of the world’s leading MOOCs, Coursera and Udacity, earned venture capital in the amounts of $22 million and $15 million, respectively. Educators, politicians, and yes, librarians, are taking note of this disruptive educational technology trend, and the future it holds for the training for our future scientists, doctors, nurses, engineers, and others entering STEM fields. I remember back my library school days, how librarians were fearing and struggling to redefine their role in “search” in the Age of Google, and now, they face another challenge, finding their role aiding professors and students in the Era of the MOOC.

ACADEMIC LIBRARIANS AND STEM PROFESSORS: A MOOC PAIRING?

The Association of College & Research Libraries (ACRL), a division of the American Library Association, presented “MOOCs & Librarians” on February 17th,  featuring talks by Panelists Valerie Hill, PhD (LISD Library Media Specialist, Adjunct Instructor, TWU School of Library and Information Studies), Ilene Frank, MLS (Adjunct Associate Professor, University of Maryland University College), Michelle Keba (MS in Information Science (Distance and Instructional Services Librarian, Nova Southeastern University) and George Djorgovski (Caltech Professor of Astronomy). The discussion relating MOOCs and librarians focused specifically on information literacy, aiding faculty by providing quality course resources, and copyright issues.

In particular, Valerie Hill pointed out that students themselves need to be cautious about the content they create as part of their participation in a MOOC, “Information literacy today requires evaluation of content like never before, with all the user-generated content coming from the bottom-up [including students in MOOC courses], rather than top-down from a high level of traditional sources. Therefore, all this talk about MOOCs made me wonder if Massively Open Online Courses would present particular problems or particular needs for information literacy. I mean, will participants [in MOOCs] have the opportunity to evaluate and to think critically about the quality of information that they are finding and getting online, and the user-generated content that they are creating and sharing on the web?” Problems with copyright violation and plagiarism might be created due to what she described as an overwhelming “expectation of freely available information” online. This, she noted, “raises concerns about intellectual property, accuracy, authority, and certainly digital citizenship, which all of these are elements of information literacy.”

WHAT WILL THE IVY LEAGUE LOOK LIKE IN 2020?

While a graduate student and a librarian at a select university, the University of Cambridge, one of the oldest British universities, and one that prides itself on its 800 year old traditions, educating the elite, and its one-on-one tutoring system

with prestigious faculty, I did the unthinkable. I enrolled in two free, online MOOC graduate courses in partnership with the Engineering Department at Stanford University, where there were hundreds of thousands in my class, many from the poorest backgrounds in developing nations, no face-to-face contact with the faculty, and a most un-traditional, and outright “disruptive” and rebellious view of the future of higher education. The courses were “Introduction to Artificial Intelligence,” taught by Sebastian Thrun and Peter Norvig, the course that gained widespread attention and started the recent MOOC movement, and became the basis for Udacity, and “Machine Learning,” taught by Andrew Ng, who later went on to co-found Coursera. I was also a lurker (someone who is interested and registers for a course and watches a few videos but does not necessarily complete the assignments or sees it through to the end) in a few other courses. Why did I do this? As a “poor student,” at a “poor college” at Cambridge, I chose not to buy a British television license, spending my precious free time entertaining myself with Netflix and 6 minute or so MOOC video lectures on the internet.

 

Of the 160,000 or so students enrolled in the AI course, I happened to be one of the fortunate 7 percent who passed and received the certificate. Prior to enrolling in this course, I had placed second in an international competition in Artificial Intelligence competition advertised by The White House and sponsored by the Department of Defense. Despite the fact that I had done well in this competition, I had no formal education in a university’s computer science program and wanted to see if I really had what it took to be able to survive one. A great advantage of the AI course was the opportunity to learn from Sebastian Thrun who I had watched on video win the DARPA Grand Challenge. After that, to me Thrun was an instant star. At least in my mind, his name became synonymous with self-driving cars and artificial intelligence.

After doing well in the DoD contest, people were asking me, why is a librarian interested in AI? What they did not know was that three librarians from Ohio [PPT] were well on their way already—two librarians in the Cleveland area, David Newyear and Michele McNeal, and me. Our AIs were no IBM Watson. We were just solo librarians pursuing pet projects, not a corporate team. They integrated AI with the library catalog and I worked on a way to use AI as a technology that librarians and archivists can use to improve and automate search and allow a natural language interface for access to academic, government, military, and commercial databases, as well as web sites. For example, in my projet, an embodied AI agent could answer questions about astronomy, after asking something like “Find Astronomy resources for quasars,” where the key search term was quasars. I designed the AI to pull results from professional databases used by astronomers like the Smithsonian Astrophysics Observatory / NASA Astrophyics Data System (ADS-Labs) database maintained by Harvard University, and ArXiv. In practical terms, Machine Learning (ML) as a specific subset of AI uses a variety of algorithms through Bayesian analysis, logistic and logarithmic regression, and other statistical means to improve performance by supervised, semi-supervised, or unsupervised learning behavior. That is, machine learning enables not just search, but automates a process for improving results through an AI’s “learning experience.” Enrolling in the AI and Machine Learning MOOC courses not only allowed me to improve my computer skills to improve my AI but to serve as one of the world’s first embedded librarians in the recent MOOC movement. As a student I familiarized myself with the course, its various discussion and help groups (even creating my own study group in Second Life that met at Stanford’s virtual library), so that I was able to gain a foothold seeking out information literacy opportunities for librarians in this new classroom of the future.

According to James G. Mazoue, writing in Educause Review Online, “MOOCs represent a post-industrial model of teaching and learning that has the potential to undermine and replace the business model of institutions that depend on recruiting and retaining students for location-bound, proprietary forms of campus based learning.” MOOC video courses can be viewed by anyone in the world free of charge, so it makes sense that online lectures by leaders in field who often find professorships the Ivy League and other selective universities would become in high demand as top “virtual professors.” Some of these may very well gain international fame as internet-driven media elevates them in the “academic pop” culture to icon status and household names.

THE PRESIDENT’S TOP ADVISORS:
PCAST’S SUPPORT OF MOOCS FOR STEM EDUCATION

The question of the librarian’s role in the development of MOOCs, which is still an emerging educational technology, is one rooted not only in technology but also in the fundamental questions of science and technology innovation policy. MOOCs, and the librarians wanting to support them, are currently struggling to find their place in, rather than disrupting, American education.

The first MOOCs in this phenomenon were computer science courses, and many more of the MOOC courses that followed were grounded in IT and technology education (circuits and electronics, cryptography, digital signal processing, image processing, social network analysis, cybersecurity, engineering, and robotics,). A range of other MOOC course offerings have also included the physical sciences (astronomy, quantum physics, and solid state chemistry), and biological and health sciences (molecular biology, genetics, health informatics, clinical and public health service, neurology, ADHD, pharmacy, and pharmacology), just to name a few.

On November 26, 2012, the National Academies of Science hosted The President’s Council of Advisors on Science and Technology (PCAST) for a meeting of more than an hour discussing “MOOCs & STEM Education.” Speakers included Sebastian Thrun (CEO, Udacity), Daphne Koller (Co-founder, Coursera; Professor, Stanford University), Anant Agarwal (President, edEx; Professor, MIT), and Frank DiGiovanni, Director (Training Readiness and Strategy, Department of Defense). Thrun pointed out three key areas needing development: technology, pedagogy, and assessment. At Udacity, he said, courses are “designed to be challenges,” not lectures, and the amount of data generated from these assessments can be evaluated using “massively using machine learning” at work behind the scenes. This approach, he said, dispels “the medieval set of myths” guiding teacher efficacy and student outcomes, and replaces it with evidence-based, “modern, data-driven” educational methodologies that may be the instruments responsible for a “fundamental transformation of education” itself.

Daphne Koller emphasized “breaking away from the one-size-fits-all model” and focusing methods for personalization and self-pacing in STEM courses where students “exhibit radically different learning patterns.” Kohler referenced earlier work by Benajmin Bloom who found that personalization can achieve “two standard deviations” of improvement in student performance [Bloom, B. (1984). "The 2 Sigma Problem: The Search for Methods of Group Instruction as Effective as One-to-One Tutoring", Educational Researcher, 13:6(4-16). (PDF)]. Yet, data-driven evaluation on a massive scale combined with effective technologies, may help fulfill the goals put forth by Bloom’s in his assumptions about the importance of personalization in education. Automated grading capabilities have advanced so that they are no longer limited to True/False and Multiple Choice, Kohler said, and include complex essays, computer programs, scientific and financial systems modeling, spreadsheets, and mathematical equations. Kohler compared the transformative effect of automation and analysis of “big data” in biology to what is happening now due to MOOCs in education. Additionally, she noted, the importance of student socialization to learn through peer grading and peer teaching. The panel emphasized that evaluation of data is just getting started, but that it may prove to be a treasure trove of information to help lead to productive changes in STEM education. 

Moderator, Eric Lander (co-Chair of PCAST, professor at Harvard and MIT, and former leader in the Human Genome Project), was apparently so enamored with the influence of MOOCs on STEM technology after this meeting, that he subsequently decided to take part himself in as a professor in a course called “7.00x: Introduction to Biology - the Secret of Life.” An MITx certificate will be issued to students who successfully complete the course and their faces will become part of a DNA photomosaic, a remake of the February 15, 2001 cover of Nature. According to another article in The New York Times, “Anant Agarwal, the former director of M.I.T.’s artificial intelligence lab, [who] is now president of edX, a nonprofit MOOC that M.I.T. and Harvard are jointly building. Agarwal [said]…some 155,000 students from around the world have taken edX’s first course: an M.I.T. intro class on circuits. 'That is greater than the total number of M.I.T. alumni in its 150-year history.’”

 By January 15, 2013, an article in Techcrunch reported that, “Today, the largest university system in the world, the California State University system, announced a pilot for $150 lower-division online courses at one of its campuses...Fed up, Gov. Jerry Brown has given his blessing to popular online course platform, Udacity, to partner with San Jose State University for the ultra-low cost online lower-division and remedial classes.” One month later, the American Council on Education approved several Coursera courses for credit.

DRAWBACKS TO MOOCS & THE CHALLENGES AHEAD

Given the advantages of MOOCs in opening new avenues to STEM education to economically disadvantaged students, career changers, and poorly performing students to help create a bridge for filling badly needed STEM jobs in the economy, what drawbacks and challenges need to be considered before deviating so far from the traditional models of education that is impossible to find our way back?

For some, MOOCs represent a threat to the educational establishment, and have provoked some educational administrators like Joseph E. Aoun, president of Northeastern University, to say that because of them we are “witnessing the end of higher education as we know it.” The fundamental business model underlying the third major portion of the economy, the university (the others being government and industry), itself may be affected. He said:

“Most significantly, MOOCs are causing higher education to shift from a vertically integrated model to a horizontally integrated one. For centuries, higher education has been a vertical enterprise: Its core functions — knowledge creation, teaching, testing, and credentialing — all have been housed within colleges and universities. MOOCs disrupt this model by decoupling teaching and learning from the campus on a mass scale …. Before long, higher education will look very different than it does today. Vertically integrated universities will continue to exist, but they’ll be joined by a variety of horizontally integrated competitors with the ability to perform the same core functions for many more people. In short, the monopoly that colleges and universities have on advanced learning and degree granting will be dismantled.”

For those who argue that this is not actually a threat and is merely the path for those unable to take a traditional route for their education, guess again. Thrun reported that the traditionally enrolled Stanford students in his AI course actually preferred the online format over his in-person lectures and performed better in that format. The reason why? They could rewind him. Despite the lack of human contact with their professors, both professors and students felt a profound connection to one another as a result of taking part in the MOOC, as demonstrated by the many thanks you and letters of appreciation on how the classes changed their lives. The professors were also affected by their impact and the hope that their students would go on to use what they learned to make advances, like Andrew Ng’s heartfelt final thank you to his first group of MOOC students.

Accreditation is one of the areas proving to be a sticking point. For the most part this might be left up to individual schools to pursue, and possibly where the future of the MOOC challenge will grind to a halt. In order to aid the universities, panelists put forth a formal request to PCAST for their assistance in speeding and simplifying accreditation procedures.

Another problematic area is student identification when taking tests. Where formal credit is earned, some universities require their MOOC students to visit a testing station in person.

The Chronicle of Higher Education reported that course quality caused Udacity to cancel one of its courses and that Coursera reportedly had dozens of cases of plagiarism. Quality control through all stages of course development and plagiarism detection software should help with this.

Finally, access to journal articles, the “meat and potatoes” staple for all science, technology, and medical students, will hinge on ready access to timely and relevant study materials. The open access mandate by funders like the NIH will help, but many more articles may be out of reach unless libraries are able to negotiate solutions with vendors and their consortia partners.

The questions librarians should be thinking about is in what ways can we leverage our skills and build tools to fit the values of personalization, “big data” modeling and analysis, as well as peer socialization and networking for the MOOC environment? Would an embedded librarian work in the future for the commercial MOOC vendor or would this be part of an Emerging Technologies Librarian’s duties at a participating university? Would libraries actually benefit financially from MOOCs when considering the added costs for access to consortial databases with copyrighted material, or would library administration need to make yet more cuts to their budget elsewhere to make room for MOOCs? Price and copyright negations with vendors are another consideration. Libraries are often charged by the number of enrolled students, how would that work if they suddenly had hundreds of thousands of students paying for a single course credit? How willing and what system will scientific journals use, for example, to allow copyright use for a selection of individual articles for a course reserve list?

Perhaps one of the best lectures I have heard on MOOC integration into the university curricula from the university administration's perspective was given by Lynne O’Brien, Director of Academic Technology & Instructional Services in Perkins Library at Duke University (slides below).

 

If a MOOC is indeed introduced into the physical classroom, where embedded librarians already are, they will need at least some level of exposure to the technology and the theoretical and pedagogical reasoning for its use in order to be able to facilitate the faculty and students. But the problem is it is so new that most librarians know little about it. Lenore R. Wilkas, M. L. S., is Associate Researcher at the Harriet H. Werley Center for Nursing Research and Evaluation and Project Librarian of the Knowledge-Based Nursing Initiative at the University of Wisconsin-Milwaukee’s College of Nursing. Her concern relates to the ability of librarians to network with one another as they face this unfamiliar territory. “Our nursing school has begun the conversation,” she said, “and I am the research librarian at the school who may also be asked to support the curriculum, so I would be interested in the comments and experiences of those already involved.” Wikas is not alone in her desire to learn more and to network about this topic. Several online webinars, like the one put on by the ACRL, are being offered. “MOOCs and Libraries: Massive Opportunity or Overwhelming Challenge” was a live webcast event offered through OCLC Research and the University of Pennsylvania Libraries on March 18-19, and a summary of was it featured on The Chronicle of Higher Education website a few days ago. Archived recordings from the event are available online. A social network or listserv group are technologies most librarians are familiar with and use, and so it seems the gap might be filled through one or more of these handled through professional associations and “a few good librarians” willing to step forth and volunteer to moderate them.