Frankly, I think that most of the time, their methods are plenty rigorous. The only difference between them and research engineering labs is that instead of writing a paper, they produce a television segment.
For some stuff, the "Can a snow plow driving down the road push enough air to flip a passing car over?" had so much wrong with it that I didn't finish the rest of the episodes on my tivo before I moved.
They tested on a dry, flat surface without any hint of ice or snow. As snow plows are not used in summer, this was a pretty bad error. A runway is flat, a road is not. I assume they were going for local location for cost and just wanted the big crash at the end as opposed to showing how dangerous a snow plow is. Coefficient of friction is an amazing thing.
Oh, their methods are quite good most of the time. The point is much rather that we know so much about how things move that if all you wanted to do, was find out whether it’s possible to bend bullets you likely wouldn’t do an experiment but break out the calculator. An experiment about this particular question can’t tell us anything new.
