<mathGeek> The Associated Press (AP) quotes (evidently without analysis) numbers from National Coalition of the Homeless.  They claim that there are 3.5 million homeless in the US.  Now, I don't know how many there are, and it has little to do with innumeracy.  But they then claim that their are 3,000 deaths of homeless people every year.  The focus of the article was on the tragedy of homeless deaths.  To me, even one homeless person having to die alone, and on the street is too many.  But here's where the math comes in...  3,000 deaths from 3.5 million homeless gives a death rate of 0.0857%.  That seemed suspiciously low.  A quick check of the CIA World Fact Book states that the death rate for the entire US is 8.34 deaths per thousand, or a rate of 0.834%.  So, if the numbers are correct, you are 10 times LESS likely to die if you are homeless. 

This seems highly unlikely to me, as homelessness seems, on first blush, to be less healthy than having a place to live.  However, it may be that homeless people tend to be young and healthy.  Or it could be that the at least one of the numbers given by the National Coalition of the Homeless is incorrect.  I'm guessing that the number of deaths is fairly well documented, while the number of homeless must be determined statistically.  If that's the case, a more likely number of homeless in the US would be around 350,000, assuming a mortality rate approximately equal to the rest of the population.  Once again, there may be many factors that influence the actual number, but a very quick 'reality-check' on the numbers reveals the possibility of very serious numerical errors.  (It looks like the NCH is overstating the number of homeless by 3.15 million.)  Homelessness is a very real problem, but drastically inflating the size of the problem for attention does nobody any favors.  Once people BS-meters are triggered, then it's hard to take anything else said seriously, thus damaging your cause. (Hat Tip: Wall Street Journal Online)</mathGeek>