"There are lies, damn lies and statistics." -- Benjamin Disraeli
There is a bit of a kerfuffle now regarding the number of deaths that occurred as a result of hurricanes Maria that hit Puerto Rico a year ago. President Trump has tweeted that when he was on the island the number was 6 to 18 ... but did acknowledge that the number later went up after he left. (The first official number of direct deaths issued by the Puerto Rican government was 64.)
However, enter stage left, politics. Sometime earlier this year the Puerto Rican government suspiciously commissioned George Washington University to investigate this matter further to see how many direct and indirect deaths there actually were. After some grinding of wheels, these statisticians came back with 3,000 deaths (2,975 to be exact ... which obviously sounds more definitive.)
But this number is artificial. It cannot be supported with names of the dead and their death certificates. It is justified only with statistics of how many people may have died directly AND indirectly as a result of circumstances surrounding Maria.
The media madness that sprang up over the disconnect between Trump's number and this latest estimate has caused me to look at the "scientific" process employed at GW University. Its method was as follows:
- Look at the number of Puerto Rican deaths in the 6-month period from September, 2017 to February, 2018.
- Compare this to the number of deaths in the same period for the two prior years.
- After some unknown adjustments, the difference must be the deaths directly and indirectly attributed to Maria.
Now, I am not an university professor, but several things bother me with this approach:
- Why a six month period? Why not 2, 4, 8 or even 12 months? What if the winter flu season in any of the sampled years skewed the results? Or any other unusual mortality rates?
- Why are deaths compared to just two years? Why not just one year or 5 years? Are changing demographics accounted for?
Bottom line, this approach is open to too much subjective judgment ft my taste. I would prefer a counting of death certificates in which the cause of death is directly or even indirectly attributable to hurricane Maria. Let's bury the politics.
