Analyzing block randomized studies: the example of the Jersey City drug market analysis experiment

Objectives

While block randomized designs have become more common in place-based policing studies, there has been relatively little discussion of the assumptions employed and their implications for statistical analysis. Our paper seeks to illustrate these assumptions, and controversy regarding statistical approaches, in the context of one of the first block randomized studies in criminal justice—the Jersey City Drug Market Analysis Project (DMAP).

Methods

Using DMAP data, we show that there are multiple approaches that can be used in analyzing block randomized designs, and that those approaches will yield differing estimates of statistical significance. We develop outcomes using both models with and without interaction, and utilizing both Type I and Type III sums-of-squares approaches. We also examine the impacts of using randomization inference, an approach for estimating p values not based on approximations using normal distribution theory, to adjust for possible small N biases in estimating standard errors.

Results

The assumptions used for identifying the analytic approach produce a comparatively wide range of p values for the main DMAP program impacts on hot spots. Nonetheless, the overall conclusions drawn from our re-analysis remain consistent with the original analyses, albeit with more caution. Results were similar to the original analyses under different specifications supporting the identification of diffusion of benefits effects to nearby areas.

Conclusions

The major contribution of our article is to clarify statistical modeling in unbalanced block randomized studies. The introduction of blocking adds complexity to the models that are estimated, and care must be taken when including interaction effects in models, whether they are ANOVA models or regression models. Researchers need to recognize this complexity and provide transparent and alternative estimates of study outcomes.

     

Gender,General Strain Theory,Negative Emotions,and Disordered Eating

Much of the prior work on General Strain Theory (GST) has focused on how strain and negative emotions interrelate to produce criminal—especially violent—activity. Very little research has extended GST to examine other types of non-criminal, negative behavior, such as self-harming behaviors associated with disordered eating, a traditionally female-specific self-directed outcome. Using a sample of 338 young adults (54% female, 95% white), this article applies GST to disordered eating by examining how strain and negative emotions relate to this particular outcome across gender. Findings indicate that two types of strain measures predict depressive symptoms among males and females, that inequitable strainful experiences relate to disordered eating among females but not males, that depressive symptoms but not anger increase disordered eating for both males and females, and that membership in Greek organizations (sororities or fraternities) is associated with disordered eating but only for males. Implications for theory and directions for future research are highlighted.     

Campbell Collaboration systematic reviews and the Journal of Experimental Criminology: Reflections on the last 20 years

Journal of Experimental Criminology -     

Criminal career progression among serious youthful offenders in Australia

An important question from research on criminal careers is whether the processes that generate participation in criminal activity are similar to those that drive offending frequency among those who have initiated. This article considers basic demographic correlates asking whether those factors are associated with both initial and sustained early adulthood criminality in Australia. Three findings emerged from the study: (1) the offenders exhibited high levels of criminal activity during early adulthood; (2) males were more likely than females to offend and offend at higher rates as adults; and (3) Indigenous youth were more likely than non-Indigenous youth to offend and offend at higher rates as adults. Overall, the study results showed that basic correlates of crime were linked to both participation and frequency of offending in early adulthood—even within a sample of serious offenders. The article addresses the theoretical implications of the findings and directions for future research.     

Truancy Intervention Reduces Crime: Results from a Randomized Field Trial

Educational attainment is a fundamental cornerstone to success throughout the life-course. As a result, ensuring that young people remain in school and are not truant is critical. Although the importance of truancy as a risk factor for many adverse outcomes, including crime, has been well-documented, much less methodologically rigorous work has been undertaken to evaluate potentially promising prevention and intervention strategies. This paper uses a randomized field trial method to test how a partnership between police and schools targeting truancy impacts offending in a sample of high-risk truanting young people. We find that the truancy intervention reduces offending and we discuss the implications for practice and directions for future research.     

Reporting guidelines for social and psychological intervention trials: considering the special context of crime and justice interventions

We comment on Grant and colleagues’ proposal to develop an extension to the CONSORT statement for social interventions, and propose that crime and justice interventions should be considered a special case, due to the added complexities of reporting consent, coercion and harm in criminal justice experiments.     

Policing Australia’s ‘heroin drought’: using an agent-based model to simulate alternative outcomes

Our paper examines how street-level drug markets adapt to a macro-level disruption to the supply of heroin, under three experimental conditions of street-level drug law enforcement: random patrol, hot-spot policing and problem-orientated policing. We utilize an agent-based model to explore the relative impact of abstractions of these three law enforcement strategies after simulating an ‘external shock’ to the supply of heroin to the street-level drug market. We use 3 years of data, which include the period of the ‘heroin drought’ in Melbourne (Australia) that commenced in late 2000 and early 2001, to measure changes in a selected range of crime and harm indicators under the three policing conditions. Our results show that macro-level disruptions to drug supply have a limited impact on street-level market dynamics when there is a ready replacement drug. By contrast, street-level police interventions are shown to vary in their capacity to alter drug market dynamics. Importantly, our laboratory abstraction of problem-orientated policing is shown to be the optimal strategy to disrupt street-level injecting-drug markets, reduce crimes and minimize harm, regardless of the type of drug being supplied to the market.

The Forward Specialization Coefficient: Distributional Properties and Subgroup Differences

For both public policy and theoretical reasons, criminologists have been interested in the degree to which criminal offenders specialize in particular crimes. Traditionally, offense specialization has been measured with the forward specialization coefficient (FSC). Recently, the FSC has been criticized for being interpretationally obtuse and having no known sampling distribution. In this paper we examine both the interpretational and the statistical properties of the FSC. We conclude that (1) it has an intuitive interpretation that is no less useful than either a standard correlation coefficient or its competitors, (2) its sampling distribution is approximately normal, and (3) the conventional formula for the estimated standard error of the FSC may underestimate the true standard error in some circumstances. With these results behind us, we propose and illustrate both a parametric statistical test for the difference between two independent FSCs and two nonparametric alternatives.