Really, I have to remind myself often that most people (even slashdot nerds) are simplistic binary thinkers. They latch onto a certain way of looking at something (ooh, I'm all sciency!) and try to hit everything with the same hammer.
What we are having here is not a scientific study. Nor is it a debate. What we are having here is a discussion. This being a discussion forum. When I bring up something anecdotal, it doesn't mean I am basing a decision or opinion purely on that anecdote. It just means I find it interesting and possibly indicative of a truth. I look at all sorts of anecdotes around me, I look at scientific studies, I look at statistical reports (statistics alone are not science), and I also like to use logical inference when I examine ideas (again, this is not science). When a broad preponderance of different types of evidence points me in a certain direction, I tend to give it credence. It's not an absolute, any more than science itself is a set of absolutes.
Here's an example of empiricism: every damn person on the planet knows that if you want to strengthen a muscle, you have to exercise it, and in general the more intense the exercise, the greater the gains. Of course there are limits after which too much exercise will be counterproductive, blah blah blah, but the core truth is there. Don't need a scientific study to prove that. It is part and parcel of everyone's experience, just like everyone knows you have to breathe in order to stay alive. When the variables get more complex of course it's not always so clear, but one has to be a particularly obtuse person not to agree that in general exercise leads to better health. By logical inference, better health would obviously lead to the likelihood of living longer. Where scientific studies help is in identifying just what the upper and lower bounds are to these benefits. But I'm not here to do your work for you. I'm just shooting the breeze on a Sunday afternoon. If you want that sort of information, complete with box plots, scatter charts, and explanations about sample size and selection bias, well you have access to the same information I do. Get to it.