So, I think the real answer is that there are many different ways in which scientific research is done. Although the "scientific method" as outlined in the 7th grade science book may be slightly problematic, real scientists take account of how far their results explain.
Example: A researcher was testing the hypothesis that hummingbirds produce a certain noise with their tail feathers during a ballistic dive meant to show off in front of females. He found that four sounds were made during this dive, which corresponded to changes in the bird's posture (wing's extended, wings folded, etc.). The timing of the noises corresponded to the timing of the different postures. The noise he was interested in was the third note. This note was produced during a phase where the bird expanded it's tail feathers into a fan. He wanted to know if the tail feathers produced the noise when they were extended.
The birds produced the noise 100% of the time with all of their feathers.
He cut off the first tail feather, and the birds produced the noise 0% of the time.
He cut off the second tail feather, and the birds produced the noise 10% of the time.
He cut off the third tail feather, and the birds produced the noise 95% of the time.
He cut off the fourth tail feather, and the birds produced the noise 100% of the time.
Now from this you might conclude that the birds used their first two feathers to produce the noise. But he did not do so. Rather, he took the first two feathers off the bird's body, and moved them through the air at the same speed that the bird was moving. He found that the feathers produced a sound at a frequency very close to the observed frequency.
At this point he concluded that the feathers produced the sound in the air. But the first question he got was from a skeptical scientist, asking him how he could not rule out the possibility that the sound was vocalized, by for example removing the bird's vocal cords. Apparently even all his work was not enough to reach his conclusion, since there was still one logical possibility left desptie the strong evidence against that conclusion.
™ Version 4.2
