Is the Curve Relating Temperature to Aggression Linear or Curvilinear? A Response to Bell (2005) and to Cohn and Rotton (2005) Brad J. Bushman University of Michigan Morgan C. Wang University of Central Florida Craig A. Anderson Iowa State University P. Bell (2005) recommended examining the relationship between temperature and assaults during the hottest times of day and during the hottest months of the year. The authors’ analyses of these data show a linear rather than inverted U-shaped relationship between temperature and assault during the hottest times of day and in the hottest months of the year. E. Cohn and J. Rotton (2005) recommended analyzing the 6 hr with the highest assaults versus the 6 hr with the lowest assaults. During high assault periods, there is a strong positive linear relationship between temperature and assault. During low assault periods, there is no relationship between temperature and assaults. Assaults and other violent crimes might decrease when temperatures are very hot, but the Minneapolis data set does not allow for testing of this hypothesis because Minneapolis is too cold. Keywords: aggression, assault, heat, hot, temperature In this article, we address the concerns raised by Bell (2005) and by Cohn and Rotton (2005) about our reanalysis of Cohn and Rotton’s Minneapolis data set. The data set contains am- bient temperatures and assault rates received by the Minneap- olis Police Department in 1987 and 1988. Cohn and Rotton plotted assault against temperature and found an inverted U-shaped curve, with assaults peaking at about 75 °F (23.89 °C). We believe this curve is misleading because it fails to take into account time of day. Both time of day and assaults are strongly correlated with temperature, but in opposite directions. Assaults are highest in the late evening and early morning hours when temperatures are coolest, whereas temperatures are high- est in the afternoon hours. Furthermore, a large portion of the time of day effect on assaults is likely due to the differential types of activities and environments in which people typically engage at different times of the day. Bell (2005) recommends examining the relationship between temperature and assaults only during the time of day when temperatures are hottest. If there is a downturn in assaults, it should occur when temperatures are hottest. In this data set, temperatures are hottest between 12:00 and 2:59 p.m. (see Figure 2 in Bushman, Wang, & Anderson, 2005). Figure 1 shows the relationship between temperature and assault during the hottest time period. As can be seen in Figure 1, the rela- tionship between temperature and assault is linear. There is no sharp downturn when temperatures are hot. In addition, if one examines the quadratic temperature regression coefficients for the 7 days of the week during the hottest time period (i.e., 12:00 p.m. to 2:59 p.m.), only three of the seven regression coeffi- cients are negative, and none of the three negative regression coefficients is statistically significant (see Table 5 in Cohn & Rotton, 1997). Bell (2005) also suggests that the downturn in aggression during the hottest time of the day should be most pronounced during the hottest months of the year. As can be seen in Figure 2, the three hottest months of the year in Minneapolis are June, July, and August. Figure 3 shows that the relationship between temperature and assault is linear during the hottest time period (i.e., 12:00 to 2:59 p.m.) in the three hottest months (i.e., June, July, August). The curve does not have an inverted U shape. Although Bell’s (2005) hypotheses make good theoretical sense, the data do not support them. The relationship between tempera- ture and assault is not curvilinear during the hottest time of the day, even if one examines only the hottest months of the year. In a hotter city, like Dallas, Texas, Bell’s hypotheses might be sup- ported by the data. Cohn and Rotton’s (2005) reply repeats the claim from their earlier article (Cohn & Rotton, 1997, Footnote 2) that removing month from the statistical model had little impact on the results and that the overall relation shows an important downturn at the highest temperatures. However, various reanalyses clearly contra- dict this claim (e.g., Anderson, Anderson, Dorr, DeNeve, & Flana- gan, 2000). Cohn and Rotton (2005) are correct in pointing out that the number of hours in the high and low assault periods is not the same in Figure 3 of Bushman et al. (2005). The high assault time period contains 6 hr (9:00 p.m. to 2:59 a.m. the next day), Brad J. Bushman, Institute for Social Research; Morgan C. Wang, Department of Statistics and Actuarial Science, University of Central Florida; Craig A. Anderson, Department of Psychology, Iowa State University. Correspondence concerning this article should be addressed to Brad J. Bushman, University of Michigan, Institute for Social Research, 426 Thompson Street, Ann Arbor, MI 48106. E-mail: bbushman@umich.edu Journal of Personality and Social Psychology Copyright 2005 by the American Psychological Association 2005, Vol. 89, No. 1, 74 –77 0022-3514/05/$12.00 DOI: 10.1037/0022-3514.89.1.74 74