Although it is generally acknowledged that benzene causes leukemia, especially acute myeloid leukemia, considerable divergences persist in the assessment of the leukemia risk due to occupational low-level benzene exposure. Specifically, the risk for vehicle mechanics is considered by some authors as being nondetectable with epidemiologic methods, whereas others calculated that the incidence rate of leukemia (all types) in vehicle mechanics is increased more than 60 times. The purpose of this review is to examine the publications on this topic in light of criteria for causal inference and to discuss the possible role of bias, confounding factors, and chance. The results of this analysis reveal that there are surprisingly few epidemiologic observations supporting an increased incidence of leukemia in vehicle mechanics. Apparently, publications suggesting a leukemogenic effect of low-level benzene exposure in garage mechanics are more often quoted than their negative counterparts, although they are not better designed.