Originally Posted by
Mistral
Ah, lies, damn lies, and statistics. I would tip my hat to you, had I a hat to tip. Why are death rates from automobiles so high in America? Because everyone and their mother owns at least one car, with a decent chance of owning two. 10% of an entire population is still larger than 100% of 5% of that population, even if the former suggests a lower risk ratio than the latter. Is the number of people consistently abusing drugs on the magnitude of half of those who own cars? Also, how does this chart define drug deaths? Is it simply a matter of overdosing, or more specifically overdose in isolation? Does this chart also include in drug-related fatalities other complications extending from prolonged drug abuse, which vary based on the drug of choice, or are these lumped into categories such as "cardiovascular diseases," "malignant neoplasms," "respiratory infections," "Lower respiratory tract infections," "respiratory diseases," "Chronic obstructive pulmonary disease," and so forth. COPD alone classifies smoking as its primary risk factor - 80% to 90% of COPD cases occur in heavy smokers, and that's a legal drug. In fact, and purely to illustrate the permissive nature of definitions, a sufficiently obstructive individual could classify a knife in the heart as "acute cardiac failure," though I, charitable as I am, do doubt that the original codifiers of the data used in this chart were quite that oblique. For that matter, there's another good way to lie with statistics - don't those last four groups I listed look a bit...similar? Splitting higher-ranked groups and lumping lower-ranked groups (which aren't listed) is a good way to push other entries down the rank structure, especially if you then extract that ranking and present it in obnoxiously bold, centered, high-size fonts as gospel.
You should think twice about the biases of your sources, and how they may affect the data those sources choose to present. Even that which at first blush may appear to be raw numbers may in fact have been thoroughly massaged to make it more suitable for the intended message, and this is by no measure the raw numbers you wished to present.