To assess the strength of the association between blood lead and children's IQ, a meta-analysis of the studies examining the relationship in school age children was performed. Emphasis was given to the size of the effect, since that allows comparisons that are informative about potential confounding and effect modifiers. Sensitivity analyses were also performed. A highly significant association was found between lead exposure and children's IQ (P < 0.001). An increase in blood lead from 10 to 20 micrograms/dl was associated with a decrease of 2.6 IQ points in the meta-analysis. This result was robust to inclusion or exclusion of the strongest individual studies and to relaxing the age requirements (school age children) of the meta-analysis. Adding eight studies with effect estimates of 0 would still leave a significant association with blood lead (P < 0.01). There was no evidence that the effect was limited to disadvantaged children and there was a suggestion of the opposite. The studies with mean blood lead levels of 15 micrograms/dl or lower in their sample had higher estimated blood lead slopes, suggesting that a threshold at 10 micrograms/dl is implausible. The study with the lowest mean blood lead level was examined using nonparametric smoothing. It showed no evidence of a threshold down to blood lead concentrations of 1 microgram/dl. Lead interferes with GABAergic and dopaminergic neurotransmission. It has been shown to bind to the NMDA receptor and inhibit long-term potentiation in the hippocampal region of the brain. Moreover, experimental studies have demonstrated that blood levels of 10 micrograms/dl interfere with a broad range of cognitive function in primates. Given this support, these associations in humans should be considered causal.