We conduct in this paper a meta-analysis on the value of a statistical life (VSL) in road safety. Accordingly, we determine the factors that can explain the variation in VSL estimates reported in the literature. Our base of work is composed by 401 observations extracted from a total of 64 published articles. We found that the VSL estimates are significantly sensitive to the GDP per capita, to time and space. The significance of the baseline risk and of the risk change in the whole sample raises the question of the validity of some benefit transfer approaches in this context. We retrieve also some common results in the literature; the VSL is directly impacted by the value assessment approach (revealed vs. stated preferences), by the elicitation format (willingness-to-pay vs. willingness-to-accept) and by the dimension of the safety policy (public vs. private).