Just ran across this quote: "Research into online learning suggests that it is roughly as effective as classroom learning." The TED talk in the OP sorta suggests the same, at a meta-level.
Chum in the water.
As a guy who used to do a lot of studies... Beware of that sort of research. Or more accurately beware of what conclusions you draw. When you're looking at a variable as difficult to define as "effectiveness of a class" and then consider what sort of range you'd be looking at, realize that it may not be possible to draw much of any conclusion beyond that. For instance, it wouldn't be inconceivable to say that many who would do terribly in a classroom for any number of reasons might do better with on-line learning. And also consider that many who would thrive in a classroom may in fact need the interaction that would be lacking in an on-line environment. Same with structure, planning, time, etc. So while things may "even out" in the end with a mean for the entire group being the same, some may thrive spectacularly in one environment while failing profoundly in the other with the exact opposite occurring for others. In other words, the overall may be true but it is meaningless for most individuals as there may be very good reasons for each individual to go for one over the other. So what you have is actually a more complex situation than the study anticipated and that may in fact call our for more options for learning rather than saying one is as "good" as the other. In other words... It ain't necessarily so and the study is flawed from the get-go because they really didn't look in the right places to begin with...