Boutwell pursued this extension since his previous study drew upon the total violent crime, and this metric can bias results because of “spatial autocorrelation”, or “the tendency for certain outcomes to cluster together geographically for reasons other than a causal association between those two variables”, in turn by deconstructing violent crime, the correlation between crime and BLL would further be supported since confounding variables would be minimized (3). In turn, Boutwell refined his crimes beyond traditional categories into, “...crimes involving a firearm, assault (with or without a firearm), 3) robbery (with or without a firearm), homicides and rape” (4). Several important findings were discovered with these refined groups, firstly “...increasing the percentage of elevated BLL [blood lead level] by just 1 percentage point was associated with 1.03 times greater risk for firearm crime...assault crime...a greater risk for robbery… and homicide” , additionally with a p-value = .05 these results mean Leads effect on these specific crimes is highly unlikely due to chance and therefore this relationship is less likely to be due to other factors like “spatial autocorrelation” and Leads affect is therefore significant in correlating with crime (7). Laslty, the other major finding from Boutwells second paper is that although the Lead and crime link is positive in regard to rape, it “wasn’t statistically signinficant”at a confidence level of p = .05, in which the results are not most likely due to chance.