A legal psychology lesson on correlation and causation
For the past several decades, Legal Psychology has been very interested in the role of perceptions of procedural justice in the legal system (a Google Scholar search turns up over a million results). Many studies have found that perceived fairness of the criminal justice system can influence an individual’s assessment of how legitimate criminal justice institutions (i.e., police, courts, and prisons) are, and can influence intentions of following the law or participating in the system (e.g., jury duty).
However, some scholars highlight that these studies might not accurately establish a cause-and-effect relationship between procedural justice and the studied outcome (i.e., they lack internal validity). Thus, Pina-Sanchez and Brunton-Smith (2020) attempted to address the lack of internal validity in their article “Reassessing the relationship between procedural justice and police legitimacy.” The authors examined 7 years of longitudinal data from over 1,000 justice-involved adolescents (from the Pathways to Desistance Study). The benefit of this database is that there was very little dropout (i.e., attrition), which means that you can more reliably assess changes over time and therefore have better internal validity.
Through their multiple analyses, Pina-Sanchez and Bruton-Smith did not find a causal relationship between perceptions of procedural justice and police legitimacy; past ratings of procedural justice did not predict later ratings of legitimacy, nor vice versa. Rather, they report that the correlational relationship might be caused by a third variable, such as individual changes in perceptions of police legitimacy over time. So, while the concepts might be linked, these findings did not support that one is driving the other (think back to your Psych 101 classes – correlation does not equal causation!).
Thus, the results of this study call into question the mechanism driving the relationship between procedural justice and perceived legitimacy. However, we should not entirely abandon our push to promote procedural justice within the justice system. First of all, this study was limited in sample and scope. The authors were only able to use the data that was collected starting in 2000. Thus, while the sample is large and there were a variety of questions, there were also limits in terms of how the authors could examine “procedural justice” or “police legitimacy” (i.e., their operational definition). Therefore, we should not entirely eliminate the existing research that has examined the relationship between procedural justice and a variety of constructs.
Secondly, even if we do change our thoughts about a causal relationship, it does not necessarily eliminate the importance of the correlational relationship. Specifically, if two items are correlated and we find a method of increasing one, it is possible that the method will also increase the other. Thus, research demonstrating techniques that increase procedural justice and a second construct of interest (e.g., police legitimacy), might not operate in the way we previously thought (e.g., by raising perceptions of procedural justice causing a raise in perceptions of police legitimacy), but may nevertheless result in the desired outcome (e.g., increase law abiding behavior) because the method influences the actual causal variable. Thus, we should conduct more research to investigate what the causal mechanism is, not necessarily because the existing interventions are not successful, but because we might be able to design even more successful interventions.