In this video (Veritasium), the author questions the effectiveness of teaching with videos like the Kahn academy's videos (Khan Academy and blogpost ELPCG1, entry4: Forget about Kumon!). He conducted a research to explore the problem.
In the first part of the experiment, students are shown a video presenting a scientific problem. They are then tested on this particular topic (test 1). In a second step, the students are presented with a second video explaining the theory and thereby revealing the correct answer. They are then submitted again to the same test (test 2). The results show that even after being presented with the correct answer, the number of students responding correctly is the same for test 1 and for test 2.
The assumption here is that when exposed to a scientific concept, people have preconceive ideas. They don't pay attention to the video giving a scientific explanation because they think their own idea is correct. They actually end up being reinforced that their preconceived idea is correct.
How can we tackle the problem of misconception? The students would pay more attention if their preconceived ideas were presented in a video. So in a second experiment, the author presents students with the same problem video and submit them to the pre-video test. They then show them a second video containing misconceptions about the scientific concept. The students were then tested again. In this experiment the number of correct answers was double after watching the video with the misconception.
So what happens there? When a scientific video is presented in a clear and concise manner, student believe they are learning, but they actually don't engage with the content on a deep level. They don't realize that what is presented is different from their previous knowledge. But given the opportunity to challenge their preconceived ideas, students were being engaged at a deeper level.