From the Boston Globe: "Study suggests online courses as good as classroom".
A new study compared two versions of an introductory statistics course, one taught face to face by professors and one mostly taught online with only an hour a week of face time. Researchers found students fared equally well in both formats on every measure of learning. The only difference was that the online group appeared to learn faster.
...
Taken together, the reports “don’t suggest that interactive online learning is far better than traditional forms of instruction - but even in its infancy, it does well,’’ said Lawrence Bacow, the former Tufts University president, who co-authored the first paper. “And today’s students become tomorrow’s faculty. They will have much greater comfort using these tools. This is only going to get better over time.’’
The study does come from a for-profit education firm. The reality is that most fields in academia give graduate students little or no training in effective teaching methods, and effective teaching is not incentivized for tenure-track faculty. It should be no surprise that a computer algorithm could do better than the average college instructor for a technical subject like statistics.
There remain several open questions. Is a good, live instructor together with online training better than one or the other separately? Can we make effective online materials for subjects that are not easily rendered as problem sets with tutorial content? Would we be better off shifting a greater fraction of introductory teaching to the fact-based assessment strategies that online materials can easily implement?