This certainly wouldn’t be a good factoid for the EPA and its illegal human testing program.
Among other reasons, this is junk science because the claim is based on an ecologic study, where individual exposures to coal plant emissions are unknown and assumed to be ambient averages.
Read more about the claim at Science Daily.
I wonder what other correlations we would come up with. Maybe check higher incomes versus living near coal fired power plant, number of people living in the area of the plant, number of cats that run loose in the area, number of bluebirds spotted on a chosen day. Correlations are so much fun to play with–and as by Stan, so are statistics!
There is a rather classic problem with the analysis of this data set, multiple testing. If you ask enough questions you will get p-values less than 0.05 by chance alone. In this study, 7 factors are examined. If you multiply the p-value in questions, in this case 0.019, by the number of questions at issue you get 7 x 0.019 = 0. 133 which is not statistically significant. So the results could well be due to chance.This is a claim not worth killing yourself over!