The relationship between pavement marking retroreflectivity and nighttime safety has been a topic of research for past decade or more but consistent findings have been elusive despite the intuitive nature of having bright markings. This paper builds from previous work that used Michigan retroreflectivity and crash data to demonstrate that pavement marking retroreflectivity relates to the nighttime safety on rural two-lane highways in a meaningful way. In this paper, new data from North Carolina were obtained and used. The North Carolina data were used to test the robustness of the statistical models derived from the Michigan data. Additional analyses were also explored and described in this paper. Using results from this paper, previous research, and state of the practice, recommendations and their implications are presented for safety-derived minimum retroreflectivity levels for pavement markings.