Three proposals for enriching neurobehavioral toxicology studies are discussed. First, while IQ has proven useful as a primary endpoint, such apical measures are limited: they obscure important individual differences, tend to reflect the product rather than the process of learning, sample a limited range of intelligent behaviors, and are insensitive to critical outcomes such as learning disabilities. In terms of societal disease burden, behavioral and psychiatric morbidities might be even more important than cognitive morbidities. Such endpoints warrant careful attention. Second, the models of child development can be enriched, increasing our ability both to control for confounding bias and to evaluate potential modification of neurotoxicant effects by contextual factors. While the use of the Home Observation for Measurement of the Environment (HOME scale) and other measures of family-level proximal factors was an important advance, recent sociological work demonstrates the importance of broader conceptualizations of the ecology of child development (e.g. neighborhood and community characteristics). Third, much effort has been expended in attempts to identify the behavioral signature associated with exposure to a particular neurotoxicant. Given the limited success in identifying behavioral phenotypes even for well-characterized genetic disorders (e.g. Fragile-X, Williams, Velocardiofacial syndromes), the prospects seem grim for identifying specific and relatively invariant patterns in the expression of neurotoxicant effects across diverse dosing regimens and biological and cultural settings. In part this results from the likely influence of complex, but largely unknown, patterns of effect modification on the expressions of toxicity. Efforts to define the nature of these contingencies might be more productive than continued efforts to identify behavioral phenotypes.