Nov 12
The foolish Greenie rush to generalizations
As the Greenies are now learning, even strong statistical correlations may disappear if a longer time series is used. A remarkable example from Sociology:
“The modern literature on hate crimes began with a remarkable 1933 book by Arthur Raper titled The Tragedy of Lynching. Raper assembled data on the number of lynchings each year in the South and on the price of an acre’s yield of cotton. He calculated the correlation coefficient between the two series at –0.532. In other words, when the economy was doing well, the number of lynchings was lower…. In 2001, Donald Green, Laurence McFalls, and Jennifer Smith published a paper that demolished the alleged connection between economic conditions and lynchings in Raper’s data. Raper had the misfortune of stopping his analysis in 1929. After the Great Depression hit, the price of cotton plummeted and economic conditions deteriorated, yet lynchings continued to fall. The correlation disappeared altogether when more years of data were added.”
So we must be sure to base our conclusions on ALL the data. In the Greenie case, the correlation between CO2 rise and global temperature rise stopped in 1998 — but that could have been foreseen if measurements taken in the first half of the 20th century had been considered.
