In a wide range of lecture-based introductory physics courses, for example, if you measure conceptual understanding of Newtonian mechanics before and after the course, you find that students only learn about 25% of what they didn't understand at the beginning.[1] (And students usually do pretty poorly on the pre-test.) With interactive teaching methods, you can boost that to 50%.
The problem is that students don't have to engage with the material during lectures -- so if they hold misconceptions, they misinterpret the lecture in light of them, rather than realizing their errors. This can be seen in lecture demonstrations, for example, where students don't learn anything from just watching a demo, but do learn if you force them to think about the problem and make predictions first.[2]
There's a methodology based on these concepts, Peer Instruction, which has students actively engage during class time by making predictions, explaining their reasoning to other students, and answering conceptual questions posed by the instructor. This has excellent results when tested.[3]
This is a hobby horse of mine[4], and I think we need to judge our claims about education using the evidence, not our personal experience. Particularly because we are the least capable judges of our learning: as students new to a field, we are completely unqualified to judge how well we understand the field.
[1] Hake, R. (1998). Interactive-engagement versus traditional methods: A six-thousand-student survey of mechanics test data for introductory physics courses. American Journal of Physics, 66(1), 64–74. http://doi.org/10.1119/1.18809
[2] Crouch, C. H., Fagen, A. P., Callan, J. P., & Mazur, E. (2004). Classroom demonstrations: Learning tools or entertainment? American Journal of Physics, 72(6), 835–838. http://doi.org/10.1119/1.1707018
[3] Crouch, C. H., & Mazur, E. (2001). Peer Instruction: Ten years of experience and results. American Journal of Physics, 69(9), 970–977. http://doi.org/10.1119/1.1374249
Given that almost all courses I took in college were the lecture / hw / exam model, I don't think that conclusion fits their claim.
Video games are excellent at teaching in a similar manner (the witness is very much like a math course that just happens to be made up). I see no reason why the digital medium can't have an equal or better experience. Good coursera courses often seem fairly interactive and the lecture itself isn't necessarily the focal point of the class.
For example, a good maths lecture is one where you can always have an idea of what you would do for the next step in a proof, so every step actually taken is something you can compare against what you would have taken.
I. Blah A. Sub-blah 1 B. Sub-blah 2 II. Foobar A. Sub-foobar 1
etc. If I can't determine how to organize my notes, then it's not really a lecture. It's just someone talking haphazardly about something they know way more about than I do. If they're introducing any concepts that I'm not already familiar with, I'll be lost pretty quickly.
This isn't a huge ask is it? I don't need an explicit outline written on the board for me. All I need is a little recognizable organization, and I can figure it out from there. Otherwise, I'm usually better served by reading a good textbook.
The hardest class I ever had in school wasn't hard because of the material, it was hard because there was no textbook and the lectures were extremely disorganized, just random ideas that weren't linked together in any obvious way. There's no need for that.