Analyzing the Evidence Exercises

As described in the introductory Analyzing the Evidence section, political scientists are often interested in examining statistical evidence for relationships (or associations) between variables in order to see if an argument is supported. However, disentangling a relationship (or correlation) from causation (a stronger relationship in which one variable causes a change in the other) can be difficult. The contentious debate over the effectiveness of the death penalty as a deterrent to violent crime is a prime example. Conflicting findings about its deterrent effect have even been cited by Supreme Court justices weighing the constitutionality of the death penalty in specific cases.i

The figure below, which shows the trend over time in the murder rate and the number of executions, is a prime example of the evidence that is often cited in favor of the death penalty’s effectiveness.

Executions and the Murder Rate

What is the relationship (or correlation) between executions and the murder rate?
Based on what you’ve read in the Analyzing the Evidence section in your textbook, would you say that this is a causal relationship or simply a correlation? Why?
When a number of different variables are trending in similar directions, it can be difficult to assess the causal mechanism underlying the subject of interest – in this case, the murder rate. These other variables can be called “confounding” factors because they make it difficult to precisely determine the individual effect of each.
What are some potential confounding variables in the study of the murder rate?
One technique that political scientists employ to deal with confounding variables is to gather more data. Yet studying the death penalty can be difficult because the number of executions is small (this is a sample size problem).
What types of comparisons might be made to better isolate the deterrent effect of the death penalty independent of such confounding factors?
In recent years, an even more controversial claim has been advanced with respect to the crime rate more generally – that a significant amount of the decline in crime during the 1990s can be attributed to the legalization of abortion in the 1970s.ii
When was abortion legalized?
Using the murder rate as a rough proxy of the crime rate more generally, trace the decline in crime reflected in the figure above. How long after legalization does the decline begin?
Would you argue that this is a causal relationship or simply a correlation? Why?

i Liptak, Adam. 2007. “Does Death Penalty Save Lives? A New Debate.” New York Times 18 November. (accessed 2/22/12). 

ii Donohue, John, and Steven D. Levitt. 2001. “The Impact of Legalized Abortion on Crime." Quarterly Journal of Economics 116: 379-420. 

For a more recent response to critics, see Levitt’s entry on the Freakonomics blog: (accessed 2/22/12).

Submit to Gradebook:

First Name:
Last Name:
Your Email Address:
Your Professor's Email Address: