Multi-factor causality and proxy data complicate deterrence research

Evidence

Verifying causal relations in human behavior is complicated by the multiplicity of possible causal factors. To complicate matters further, many factors correlate with each other to some degree, requiring statistical methods to estimate how much each of the various factors contributes to behavior on its own. A reliable casual theory is required, to provide guidance on which of many factors should be selected for the causal analysis. The strength of this analytic model lies in its ability to gauge the deterrent influence of the death penalty after the influence of other possible causes of murder rates are taken into account. However, controversy has arisen over its application. Some critiques are directed at the explanatory value of the causal model. This is a key issue because it determines which factors among the many possible causal ones are selected. Other concerns center on the quality and completeness of the data and whether they are measured at the individual or aggregated group level. Often, data for key factors are unavailable so researchers have to rely on secondary or proxy measures for them. For example, the unemployment rate might be used as a proxy for incentives to engage in criminal activities. Still other issues involve disagreements about the statistical methods used. Are the findings based on cross-sectional or time-series data? Are data on social conditions that are conducive to criminal behavior aggregated at the national level or for local jurisdictions? How do law enforcement and judicial practices affect the risks and consequences of crime? — Albert Bandura, Moral Disengagement, 1949

Categorization

• QUESTION: The debate over which causal factors should be selected for death-penalty analysis mirrors the chapter’s broader theme of how analytic models can bias conclusions.
• corroborated: 2 independent passages support this claim — Albert Bandura, Moral Disengagement: How People Do Harm…, 1949