Lynchings and the Economy: A Time-Series Reanalysis of Hovland and Sears (1940)

Hovland and Sears (1940) reported high correlations between economic indexes and the number of Black lynchings in the American South. We reanalyzed the Hovland-Sears data set, a classic in social psychology, by using modern time-series techniques to provide better controls for the possible artifacts of trend, seasonality, and serial dependency. Correlations between the economic indexes and number of lynchings were significant, but the magnitude was less than was reported by Hovland and Sears. A new analysis examining the relation between White lynchings and economic indexes following the procedure used by Hovland and Sears yielded positive correlations, contrary to expectation. However, when time-series techniques were applied to these data the correlations were negative and approached statistical significance. Also, Davies's (1962) temporal relative deprivation hypothesis was partially supported when lagged variables were used. Thus, this article illustrates the importance of applying time-series techniques to temporal data to control for common artifacts.