Letters to the Editor
-
Another problem...
...with the way this study is being analyzed. You cannot say that because people who fight die X amount later it's the fighting that caused them to die later. A certain type of person tends to fight or not fight; a certain type of person tends to marry another certain type of person, and they tend to do so at a certain age, and so on and so on. There is a maze of variables here, featuring an endless number of permutations, and frankly, I don't see how this information can possibly be useful, especially given the sample size.
I acknowledge that these sorts of studies cannot be carried out in anything resembling lab conditions, but it's still important to keep in mind the nature of statistics and how these sorts of studies are ideally supposed to work. Data always distort and thus never produce an exact, perfectly clear picture of anything, but the idea is to isolate variables so that you've got as little distortion as possible. The clarity of the lens through which we observe an alleged phenomenon determines how useful the information can be.
This study is so far from the ideal that one must be very, very careful in analyzing it for the purpose of gleaning anything useful. Sure, it's no lab experiment, but if, say, the FDA used standards this loose we'd probably all be dead by now.