- Candace Thille moves to Stanford
- Liberal arts college explore uses of 'blended' online learning
- Report: Robots stack up to human professors in teaching Intro Stats
- Essay on how technology and new ways of teaching could upend colleges' traditional models
- Candace Thille talks MOOCs and machine learning
Hybrid Education 2.0
What if you could teach a college course without a classroom or a professor, and lose nothing?
According to researchers at Carnegie Mellon University, there’s no "what if" about it. Earlier in the decade, Carnegie Mellon set out to design software for independent learners taking courses through the university’s Open Learning Initiative, an effort to make courses freely available to non-enrolled learners. But rather than merely making course materials available to non-students, like MIT's famous OpenCourseware project, Carnegie Mellon wanted to design courses that would respond to the individual needs of each student. It currently has courses in 12 different subjects available on its Web site, mostly in math and science.
In the process of testing the software on Carnegie Mellon students to make sure it would “do no harm” if used, the researchers found that, over a two-semester trial period, students in a traditional classroom introductory statistics course scored no better than similar students who used the open-learning program and skipped the three weekly lectures and lab period.
Carnegie Mellon is not about to replace all its professors with computer programs. But with $4 million in private grants and perhaps more to come from the federal government, the university is currently exploring how the open-learning software could be used in conjunction with classroom education to speed up the teaching and learning process -- a prospect that some involved think could help solve overcrowding in America's community colleges and realize the Obama administration's goal of boosting graduation rates.
As intriguing it was to find that a computer program could prepare students to pass tests just as well as a professor, the researchers seem more excited by a hybrid application of the open-learning program that, instead of replacing professors, tries to use them more effectively. By combining the open-learning software with two weekly 50-minute class sessions in an intro-level statistics course, they found that they could get students to learn the same amount of material in half the time.
“If they’re all getting that baseline information, [faculty] can spend that class time going deeper and doing something much more interesting, so they can really leverage that you’re an expert,” says Candace Thille, director of the Open Learning Initiative, “because right now, oftentimes the faculty expertise is wasted.”
The project’s 2008 report on the tests takes care to state that the goal of the testing was not to prove that professors are superfluous to introductory-level courses, but rather to make sure students who don’t have access to a classroom course can take an “equivalently effective alternative.” That those courses might produce indistinguishable outcomes from professor-taught classes attended by tuition-paying students is nonetheless provocative.
“At the most selective tier of colleges and universities, they have some significant interest in the existing model of residential education,” says Roger C. Schonfeld, manager of research at Ithaka S+R, the strategy arm of Ithaka, a non-profit higher-ed technology group. “And I think there’s a lot more at risk in terms of the reputation they have built up over the course of decades or centuries, that even for the many advantages that might come from new models, there may be obvious or unforeseen disadvantages they need to guard against.”
Gary Rhoades, general secretary of the American Association of University Professors, says that the intangible benefits of “contact with human beings" are still crucial to learning, and that plenty of research has said so. “So many of the effects that we seek from education… go well beyond the limited ‘learning outcomes’ that are defined for a specific subject matter in a specific class,” Rhoades says. “Those outcomes generally are connected to a narrow set of competencies focused on the specific subject matter.”
Still, the apparent quality of Carnegie Mellon’s online learning program does touch on the prickly subject of what should qualify students to receive credit for material they demonstrate having learned. “If we can demonstrate that these courses are effective at supporting these students to achieve these outcomes even without being in a formal class, why isn’t someone saying OK, work through the [online] course, and if you work through it successfully we’ll say, ‘Great, you’ve passed statistics, here’s credit!’ ” says Thille. “And that’s actually a really big question in the open educational resources movement right now.”
Thille says the current model relies on a credentialed instructor to confirm what written assessments suggest: that a student has either earned institutional credit or not. And while she does not dismiss this model, and makes clear that her institution has no plans to “do away with” professors, she does note that a logic course, currently taught to residential students for credit through the hybrid model, only involves a cursory level of instructor contact. “It’s completely taught self-paced and online, and they put a [teaching assistant] who is, in theory, responsible for being in the course,” Thille says. “Basically what they do is just monitor the grade books… The students are pretty much working through it on their own, and at some point they’re done, and then they get credit for taking logic.”
The university also plans to move a non-credit digital literacy course from its current incarnation as a class taught in clusters by teaching assistants to one taught completely online by the software’s digital tutors.
But what the Carnegie Mellon researchers are selling -- and what the U.S. government might be looking to buy -- is greater efficiency, which is the promise of the hybrid version. The university last week announced it has received $4 million from the Hewlett, Gates, and Lumina foundations to help build a version of the Open Learning Initiative specifically for community colleges, citing the White House’s proposed “American Graduation Initiative,” an effort currently under review in Congress that would invest up to $500 million in online learning projects, possibly based on the Carnegie Mellon model.
‘Reinventing Higher Education’
So what exactly is the pedagogical model Carnegie Mellon has discovered, that has inspired such faith? Essentially, it’s an online program that teaches students itself, rather than just being the medium a professor uses to teach. Furthermore, it leverages the opportunity to interact directly with a unique student -- an opportunity a professor addressing dozens of students in a lecture hall does not have.
“Studies have shown that immediate and targeted feedback leads to significant reductions in the time it takes students to achieve a desired level of performance,” the researchers wrote in a paper on the project published last year. “Distributed throughout OLI-Statistics, there are many ‘mini-tutors,’ interactive activities that give students hints and feedback as they practice individual skills,” they continue. “Each of these was carefully constructed to respond to particular mistakes and misconceptions students would likely show.”
In other words, the software acts like a private tutor, quizzing students constantly as they work through linear lessons and adjusting in accordance with how quickly they show they are grasping different concepts. Testing companies have used a similar concept to make standardized evaluations such as the GRE and the GMAT adapt to the abilities demonstrated by the test-taker, but using adaptive technology as a teaching tool is relatively novel.
The virtual tutor takes care of the basic concepts that typically dominate lectures, leaving professors open to plan the face-to-face component of the course according to what parts of the curriculum the software tells him students are picking up more slowly, and what concepts could bear reinforcement. For example, if a statistics professor notices in the data he receives from activity in the open-learning program that a great number of students struggled with the assessments the program gave while teaching conditional probability, the professor could use the class periods to hold a discussion with his students about that concept until he is confident they get it -- a preferable alternative, Thille says, to rolling through concepts didactically and hoping they stick.
“I would call that reinventing higher education,” says Thille, noting her hope that this new paradigm of course design could render lecture courses -- and lecture halls -- obsolete. “It sounds Pinky and the Brain-like, but that’s actually what I’m trying to do.”
Lectures and the classroom spaces built to accommodate them, she explains, represent academe’s first attempt to do higher education at scale after new waves of students started flooding into America’s universities following World War II. With technology having evolved to its current state, such a method is primitive, she says. “You have this poor faculty member,” Thille says, “who’s sitting there as an expert in their area trying to figure out how to transfer their expertise to this large number of students, who are all variable. And it’s a horrible task. The affordances of the technology and also the learning sciences… for the first time enable us to really think about how to scale in a much more effective way, so we truly can serve many many many more students."
Search for Jobs