Geophysical variables and behavior: LXXIII. Ubiquitous errors: a reanalysis of Anderson's (1987) "temperature and aggression"

Abstract

Reanalysis of Anderson's 1987 data showed that correlations between number of hot days and violence shrank to nonsignificance when variables indicative of the South's subculture of violence were statistically controlled. The reanalyses also indicated that Anderson's results were unique to the procedures he used to select sociodemographic controls, operationalize temperature, and combine crime rates. It is concluded that relations described in a review of research on temperature and violence are neither ubiquitous nor effects.