When might it be necessary to use an experimental approach rather than an analytical one?

Answer 1

Experimental approaches are required when the science is not fully understood.
Classical physics is well understood and nearly any problem can be solved analytically. Landing on the moon was not an experiment, it was done first time because it was solved analytically.
Life science is poorly understood and experiments are necessary to confirm expectations. New drugs often fail because even after applying all that we know they may prove ineffective or harmful.

Answer 2

When you think you've worked out all the 'bugs`
analytically, it's necessary to confirm it by experement
before commiting major resources.
That's what prototypes are all about.
There is no substitute for empirical data.

Answer 3

Always. Analytical calculations are good and fine but they ain't worth a thing without an experiment. And it is never "rather than" but always "in conjunction with".

Answer 4

I have similar project , the experimntal results needs to be very accurate , consistenet and detailed , and it will shows what is wrong on the theory