Practice Exercises

1. Complete the Interactive Exercise for Chapter 6.

2. Thomas Rotolo and Charles Tittle (2006) were puzzled by a contradictory finding about the relationship between city population size and crime rates: The results of most crosssectional studies differ from those typically obtained in longitudinal studies. To test different causal hypotheses about this relationship, they obtained data about 348 cities that had at least 25,000 residents and adequate data about crime at four time points from 1960 to 1990. Let’s review different elements of their arguments and use them to review criteria for causality.

a. Cross-sectional studies tend to find that cities with more people have a higher crime rate. What criterion for causality does this establish? Review each of the other criteria for causality and explain what noncausal bases they suggest could account for this relationship.

b. Some observers have argued that larger cities have higher rates of crime because large size leads to less social integration, which in turn leads to more crime. Which causal criterion does this explanation involve? How much more confident would you be that there is a causal effect of size on crime if you knew that this other relationship occurred also? Explain your reasoning.

c. Evidence from longitudinal studies has been more mixed, but most do not find a relationship between city size and the crime rate. What do you think could explain the lack of a longitudinal relationship despite the cross-sectional relationship? Explain.

d. Some observers have proposed that the presence of transients in large cities is what leads to higher crime rates because transients (those who are not permanent residents) are more likely to commit crimes. What causal criterion does this involve? Draw a diagram that shows your reasoning.

e. In their analysis, Rotolo and Tittle (2006) control for region because they suggest that in regions that are traditionally very urban, people may be accustomed to rapid patterns of change, whereas in newly urbanizing regions this may not be the case. What type of causal criterion would region be? What other factors like this do you think the analysis should consider? Explain your reasoning.

f. Now you can examine the Rotolo and Tittle (2006) article online (if your library subscribes to the Journal of Quantitative Criminology) and read the details.

3. Search Sociological Abstracts or another index to the social science literature for several articles on studies using any type of longitudinal design. You will be searching for article titles that use words such as longitudinal, panel, trend, or over time. How successful were the researchers in carrying out the design? What steps did the researchers who used a panel design take to minimize panel attrition? How convinced are you by those using repeated cross-sectional designs that they have identified a process of change in individuals? Did any researchers use retrospective questions? How did they defend the validity of these measures?