In a Times News Review interview, Marcus du Sautoy, our new Oxford University Professor for the Public Understanding of Science, says:
I became a mathematician and not a scientist because science often goes wrong .... You do an experiment 100 times and you get the same result. You do it for the 101st time and something different happens. No one would be my lab partner at school because all my experiments went wrong. Maybe I’m a bad choice for this job.
This set me thinking.  Maybe some high-school sciences tudents are intimidated by the thought of an experiment as "something you've got to get right".  I have seen this misapprehension carry over, even into some of our PhD students.  Carl Wieman, who has put three articles on Optimizing Science Education on this blog, has suggested that simulated experiments may sometimes be better for teaching physics.

What do you all think?