Consistently inconsistent: Examining variability in police decision-making in mental health calls using a novel noise audit approach*

Purpose: As the universe of potential response options to address mental health-related calls for service continues to grow, it becomes increasingly important to understand how police make decisions during interactions with mental health involvement. This study evaluates the levels of variability in the preferred course of action police select to address these encounters. Methods: We employ vignette-based surveys in two police agencies to evaluate the level of disagreement in the preferred categorical outcomes of calls involving people with mental illness. To understand factors that contribute to levels of variation in responses, we randomly assigned participants to one of two variations each vignette, which modifies a single element of the scenario. Results: We find extensive evidence of 'pattern noise', or variability in the preferred course of action when dealing with mental health calls for service. This disagreement is observed across jurisdictions, among both patrol and command staff, and in varied situational contexts. Conclusion: The application of noise audits in law enforcement agencies can help to quantify the level of alignment in how decisions are made among both patrol officers and by leadership. This process can facilitate efforts to align organizational decision-making.