Lurk
Happy Creepy Ass Cracka
Normally, I agree. I usually want to at least SEE the rationale behind the statements. While a relationship between A and B doesn't mean A causes B, it's both mathematically and scientifically inaccurate to dismiss it altogether. If there's a solid +1 or -1 correlation coefficient between the two items (meaning, roughly, there's a precise linear relationship between the two), it's dishonest to dismiss A as a cause of B.
As you observed with the can - there was another overlooked cause. Happens all the time - Bubonic plague was blamed on rats. It was fleas, or rather, germs they carried. Admittedly, the effect was in direct proportion to the amount of rat infestation, but the wrong reason.
But we do tests and studies all the time when we don't know the exact reasons behind an observable cause and effect. We give one group a pill, and another, a placebo and measure the differences. We do this because - well, we don't know every possible interaction, but we think we do. We do experiments this way - try something, and compare it to a trial where the outcome is known. It would be bad science to treat the two as unrelated, separate experiments.
Too few know of or understand the concept of epiphenomena.