Originally Posted by
Drwecki
YOUR DATA DON'T DO THIS . THAT"S THE WHOLE POINT. It's a null result.
You are basing conclusions on null results. That is my point that is dumb in the scientific world. You can argue all you want, but you would need to show a significant difference some how some way, not an insignificant difference. Re-read the wiki page. I'm sorry I should have known better to try to help people understand science. Anyways, keep thinking that these data falsify anything. They don't, they won't, and they never can. You don't need to wear a helmet but don't invoke science as the reason, because this is simply not how you are supposed to interpret data. You would fail any intro stats class by making your conclusions. You say you falsified something, you haven't! Non-significant differences mean "Of no consequence." THIS IS A STATISTICAL FACT!
Okay, fair enough. I shot myself in the foot by using the word 'falsification'. I did not use the correct word. This is not a falsification, a proof by counterexample. It is not a proof (or disproof in this case) at all, but I've been saying this all along.
I stand by the rest of what I said. In this case, a positive correlation was hypothesized and none was observed. To say that the null result is useless because it doesn't disprove anything is crazy. Again, the Michelson-Morley experiment produced a null result, but it was hardly useless. It may not have proved that we're not surrounded by a medium through which light travels. But based on what people understood about disturbance of a medium, the original hypothesis needed to be reevaluated.
The experiment, if the method was sound, should have produced a positive correlation, but observed none, and so the hypothesis was
less likely to be correct.
To borrow an example from the Simpsons--I sell you a rock that I say keeps bears from attacking you. You buy it and, sure enough, you are free from bear attacks. But then someone else comes along and says, "That rock doesn't work," and takes it away, and you still aren't attacked by any bears. Does this disprove the original hypothesis: that the rock was keeping bears at bay? No. It's possible hunting season started the day I took the rock away and most or all of the bears were killed. It's also possible that even though I took the rock away in March, most of the bears were still hibernating because of a late winter. But given the likelihood of either of these explanations or lack of it, it would be prudent at this point to suspend your slavish devotion to the original hypothesis until some corroborating evidence could be observed.
Again, the hypothesis here, that helmets save a significant number of cyclists' lives, isn't disproved by the null return, but given the fact that no corroborating evidence has been produced, and given the unlikely nature of alternative explanations of the data, it's time to reevaluate the original hypothesis: the idea that helmets are somehow a necessary part of bicycle safety equipment.