By 2016, ongoing police shootings of unarmed black men had increased public awareness of policing processes; chief among those was militarization. Prior researchers explored this relatively new process, but very few had systematically analyzed it. In this paper, I conceptualize, operationalize, and describe police militarization across the United States. I transformed publicly available data, the Law Enforcement and Administrative Statistics (LEMAS) dataset, into a panel of over 4,000 police departments over an eighteen-year period. I then constructed an index based on each of three dimensions of militarization: (1) increased stock of military equipment; (2) increased usage of paramilitary tactics; and (3) increased focus on military-style policing. I explain the indicators within each sub dimension, and describe their unique trend across time. Using both this Absolute LEMAS Militarization Index (ALMI) and a summary analysis of each indicator, I describe militarization between 1990 and 2007 using hierarchical linear modeling. Among key findings are three trends. First, I provide nationally representative evidence of an increasing trend in police militarization across most departments, despite statistical controls. I also found that larger departments were both more militarized and militarized faster over time than smaller departments. Finally, I found that diverse departments (either proportionately more women or more black officers) militarized at a slower rate over the period. This paper provides the theoretical and methodological basis for follow-up analyses of processes affecting militarization, including variations in crime control, organizational characteristics and racial dynamics.