A recent article by kleinerman et. al. uses an "exposure index" similar to a model for residential magnetic field exposures that we developed. Magnetic fields predicted by our model were associated with childhood leukemia (odds ratio = 2.00 in the highest exposure group, 95 percent CI: 1.03, 3.89, p for trend = 0.02) (3), while Kleinerman et. Al. were unable to find an association (odds ratio for a continuous exposure variable = 0.95, 95 percent confidence interval: 0.78, 1.16). Potential explanations for these conflicting results are worth discussing because the carcinogenicity of electric and magnetic fields (EMF) is unresolved. This letter focuses on two possible reasons.