These experiments / "experiments" tend to involve a significant factor that would invoke bs'ing. E.g. being interviewed on camera from a crowd - saying "uh" and "dunno" instead of answering is guaranteed to get your clip cut. Or in a scientific experiment: our society values knowledge. So if your amount of knowledge is being measured, a natural response would be to try to look extra good.

If they did am experiment like this, informing everyone a priori that some of the statements are bs, and that at the end, they're graded on identifying the bs as well as their knowledge, that could lead to a different outcome. At least it's one way to somewhat mitigate the desirability of appearing knowledgeable.