A re-examination of correlations between lynchings and economic indices.

HOVLAND and Sears (2) attempted to verify a causal relationship between economic frustration and aggressive behavior as expressed in lynchings by a statistical study. The statistical technique was as follows: linear trends were computed for the numbers of lynchings reported for the years 1882 to 1930, inclusive, in the United States. Two figures were available for each year, the total number of lynchings in the United States, including both colored and white victims, and the number of lynchings of Negroes, most of them occurring in fourteen Southern states. Thus two trend equations were computed. The method of least squares was used in the computations. Similar computations were performed for the total value (in dollars) of cotton produced in fourteen Southern states and for the per-acre value of the cotton. The deviations of the numbers of lynchings and of the cotton value from the figures predicted by means of the trend lines were determined. In addition, an average Ayres index was computed for each year from the monthly composite indices of economic activity reported by Ay res (i). Thereupon a number of tetrachoric correlations were obtained by using two approximation methods: