Controversy over lead's effect on children's cognition rests in part on the assumption that if such an effect exists it can be characterized by a single estimator (e.g., the same rate of decline in IQ with increasing exposure, the same neuropsychological presentation), which will be found by any study that is valid. Accordingly, efforts to resolve inconsistencies in study findings have focused almost exclusively on data analytic issues germane to bias, in particular confounding and its statistical control. Relatively little consideration has been given to the role of effect modification, i.e., the impact on effect estimation of differences in the "experimental systems" employed in human epidemiological studies. Lack of consistency in findings could be due to differences among study cohorts in exposure/toxicokinetic factors (e.g., dose, timing), differences in environmental characteristics (e.g., co-exposures, co-morbidity, developmental supports, assessment setting), or differences in the distribution of genetic characteristics that affect lead metabolism. Recent findings regarding lead's impact on the development of nervous system structure and function are consistent with the hypothesis that contextual factors affect the form in which lead toxicity is expressed and may contribute to the failure to date to identify a lead-associated "behavioral signature." Characterizing the neuropsychological effects of lead might be facilitated by greater use of a clinical "process" approach to assessment, which would permit the type of fine-grained analyses of lead-associated performance differences often employed in studies of behavioral toxicity in animal models.