The Effect of Publicized Life Sentences, Death Sentences, and Executions on Homicide
Steven Stack, Criminal Justice, Wayne State University
Research Grant, 1998
The present study addressed several neglected issues in research on the death penalty and homicide:
Publicized Life Sentences vs. Executions. Research on the social control of homicide has focused almost exclusively on executions to the neglect of life sentences as a possible form of social control. The present study weighed the relative importance of publicized life sentences against publicized executions in the control of homicide. The results of a multiple regression analysis found no evidence of a reduction in homicide following a widely publicized life sentence story. In contrast, a widely publicized execution story was associated with a dip of 3.3 homicides for Whites (but not African Americans). Publicized executions, by providing severe enough punishment, may be preferable to life sentences in efforts to control lethal violence.
Publicized Executions: Abolitionist States. A second contribution was to test two alternative theoretical explanations for dips in homicide after a publicized execution. Publicized executions may reduce homicide through deterrence or through validation of societal norms against killing and their promotion of precautionary behavior (descriptions of the horrific crimes of offenders may remind persons that they are in danger, and some may avoid dangerous people for a while). The present paper contributes to the literature by exploring the impact of publicized executions in abolitionist states. A multivariate time series analysis determined that publicized executions were unrelated to homicide in abolitionist states. The results are consistent with the view that publicized executions may reduce homicide mainly through deterrence and not other processes from the theoretical literature.
Publicized executions, by providing severe enough punishment, may be preferable to life sentences in efforts to control lethal violence.
Critique of Previous Research. A third major contribution was to explain the many discrepancies in the findings on publicized executions and homicide. There have been considerable differences in the methodology used in these studies, which might help to explain their mixed findings. The present study analyzed 385 findings contained in twenty-five research studies published between 1935 and 2000. The results of a multiple logistic regression analysis determined that studies that assume execution stories will impact homicide rates for a period of more than thirty days were 95 percent less likely to report a homicide dip than studies assuming an effect period of thirty days or fewer. Studies based on the period after 1975, an era of declining celerity or speed of execution (swift punishments are a key assumption in deterrence theory), were 58 percet less likely to report a homicide drop than their counterparts. Studies based on analyses with an uncorrected problem of statistical multicollinearity were 97 percent less likely to report homicide dips than their counterparts. The model correctly predicted 91 percent of the 385 findings. It is suggested that researchers need to build a stronger consensus on methodological issues in order to reach a firmer conclusion on the impact of publicized executions on homicide.
Methodology. Homicide data were taken from the US Public Health Service’s annual Mortality Tapes. There were 164,199 homicides in this data set. Daily counts for each of the 2,922 days in the study period were obtained. Data on highly publicized life sentence and execution stories were taken from standard databases including the Vanderbilt Television News Archive, the New York Times Index, and Facts on File. A total of seventeen executions and fifteen life sentences met the criteria for being highly publicized stories. Controls are incorporated into the relevant statistical models for factors related to homicide and include holidays and temporal factors such as day of the week and season.
- Stack, S. 2001. "Publicized Executions and the Incidence of Homicide: Methodological Sources of Contradictory Findings." In DuPont-Morales, Toni and Michael Hooper (eds), Handbook of Criminal Justice Administration, New York: Marcel Decker, pp. 355-369.