On a mailing list, someone asked: “Isn’t empirical evidence found anywhere? From anybody?”
No. Absolutely not. Empirical evidence comes from doing the experiment yourself, or as close as you can get to it. If I take the average of the guesses of the peasantry as to the length of the emperor of China’s nose, why should I expect it to be at all related to the actual length? I should go and measure it myself. Or at least get his valet drunk and ask him if I can’t actually abduct the emperor, cut off his nose, and lay it next to a ruler myself.
The only empirical evidence that you can get from asking a peasant how long the emperor of China’s nose is, is what that peasant is willing to publicly state as his guess as to the length when asked by some random person he doesn’t know who’s bothering him out of the blue…and even that’s biased to peasants that will answer you as opposed to chasing you out of their field with a hoe.
As for the links you sent, unless you’ve actually read and evaluated the original papers, you can’t have a prayer of knowing what the truth of the situation is. Even if you read the original papers you may not, since they may all be unusable and broken. Actually, the majority of the published literature is. It’s all too easy to cherry-pick a few papers to support your position, just like you can round up all the folks who’ve actually measured the length of the emperor’s nose into a stockade and demand under a hot light that they tell you the result of their measurement, but it’s so easy to assume that those guys with the overly long measurements were just trying to tell you what you wanted to hear since you were applying the light pretty close to their pasty, white faces.
Want to have an informed opinion? Go cut off the nose yourself. Or go actually run your own experiment.
Obesity rates? Set up a survey sample. Hint: the proper way to do this is to intern the population of your test area, assign them numbers, pick at random from them (and make sure you weigh any who were shot trying to escape before disposing of the bodies, in case their numbers are picked), and then strip, hogtie, and weigh those selected. Unfortunately, this is illegal in the USA unless you’re only interested in numbers from those who are Japanese or black.
Autism-vaccine link? Go vaccinate a bunch of kids, and inject the others with water. Stick ‘em in isolation bubbles so they don’t die of anything, or you’re confounded by kids killed by stuff they weren’t vaccinated with. Be sure to have a lime pit on hand to dispose of the bodies of parents trying to rescue their kids. Warning: this kind of longitudinal study takes a great deal of time. A decade or so. Again, there are legal problems with this.
That is what I know about empirical evidence.