When researchers in the last part of the twentieth century demonstrated that lead levels in children too low to produce obvious physical symptoms do great harm to their cognition and behavior, they launched a major advance in public health. Clinical symptoms of lead toxicity, including abdominal pain, clumsiness, headaches, gross behavioral changes, and worse, typically become visible in children at blood lead levels greater than or equal to 60 micrograms/deciliter . In the 1970s, the U.S. Centers for Disease Control defined acceptable blood lead level accordingly, using 60 μg/dl as the threshold for concern . In 1979, Needleman et al.  demonstrated that dentine lead levels in groups of children without clinical evidence of lead poisoning were significantly associated with measures of cognitive as well as social functioning (including IQ tests, measures of verbal and auditory processing, and school behavior measures). Extensive subsequent research produced similar findings over a wide variety of geographic locations [3–6]. A meta-analysis performed by Needleman and Gastonis in 1990 consolidated the findings of twelve of these subsequent studies into strong evidence for the fact that lead causes significantly impaired cognitive functioning in children at blood levels far below the threshold of concern of the 1970s .
Further research on the effects of pediatric lead exposure on IQ specifically has quantified the average IQ loss associated with certain changes in blood lead level. Several studies have estimated a 1 to 5 point decline in IQ for every 10 μg/dl increase in blood lead . Several others have focused on the cognitive effects of lead at the lower end of the lead exposure spectrum: Bellinger, Stiles and Needleman found an average decrease of 5.8 IQ points with a change in blood lead from 0 μg/dl to 25 μg/dl; Schwartz reported an average decrease of 2.6 IQ points with an increase from 10 μg/dl to 20 μg/dl; Lanphear et al. estimated an average decrease of 6.9 IQ points with an increase from 2.5 μg/dl to 30 μg/dl [9–11]. And others have explored the idea that the relationship between blood lead and IQ is non-linear, and that much of the damage is in fact done over the first few micrograms per deciliter. In 2003, for example, Canfield et al. found that an average blood lead increase from 1 μg/dl to 10 μg/dl was associated with 7.4 point decline in IQ . It is still unknown whether there is any level of lead exposure below which harmful effects do not occur; current CDC guidelines set a threshold of 10 micrograms per deciliter. Throughout this article, we will refer to this level of 10 μg/dl as a threshold for concern, and levels above this as "elevated," without implying that a biological mechanism corresponds to this arbitrary cutoff point, and particularly without implying that levels below this are harmless.
The influences of lead exposure on cognitive and social functioning have been noticed with a wide variety of outcome variables. In a long-term follow-up with the participants in his 1979 study, Needleman and colleagues found that childhood lead exposure was associated, eleven years later, with lower class standing, increased absenteeism, lower vocabulary and grammatical reasoning scores, and more self-reported delinquent behavior, as well as a markedly higher risk of high school dropout and reading disability . Another set of children with elevated blood lead levels scored higher on measures of attention, aggression, and delinquency after other covariates were accounted for . In another group of children, prenatal and postnatal lead exposure was associated with parent and self-reports of delinquency . In adulthood, the social effects of lead exposure can manifest themselves in criminal behavior: the Philadelphia cohort of the Collaborative Perinatal Project found that childhood lead exposure was the strongest predictor of adult criminality . Several ecological investigations have found societal lead levels to be associated with levels of violent crime . Thus the effects of lead exposure can be noticed in variables less precisely quantifiable than IQ. We see that lead has the potential to impact individuals' functioning across a wide variety of domains, with the ultimate effect of impairing the functioning of society.
Gradually, the research findings of the late twentieth century were translated into action. In 1978, lead paint was banned from residential use in the U.S. . The U.S. Environmental Protection Agency issued its first regulation of leaded gasoline in 1973, which initiated a phase-down process for the next several years; in 1996, the Clean Air Act was amended with an official ban of the sale of leaded fuel for use in on-road vehicles . In 1986, the Safe Drinking Water Act outlawed the repair or installation of leaded pipes, fixtures, solders, and fluxes . By the early 1990s, the percentage of children with blood lead levels above the threshold of 10 μg/dl had fallen dramatically. Whereas 88.2% of children in the United States had blood levels exceeding that threshold between 1976 and 1980, that figure had fallen to 1.6% by 2002 .
This story represents an enormous public health success. Has the mission been successfully accomplished? As a USA Today article declared of lead paint in 2007, "It's banned, but not gone" . According to estimates of the American Academy of Pediatrics in 2005, one in four children lives in housing that contains deteriorating lead paint . According to estimates of the National Center for Healthy Housing, renovation, remodeling, and repainting of older housing stock exposes approximately 1.1 million children in the United States to lead annually . And worldwide, children are still being exposed to lead in even larger numbers, often as a result of industrialization in formerly "underdeveloped" countries .
How noticeable are the effects of lead on academic performance? In this study we use recent data on lead exposure and school achievement in counties across New York State to explore the continuing effects of lead upon children.
In addition to the sizeable body of research that has examined the influences of lead exposure on IQ, more recent studies have come to focus more on its effects on school achievement. Some have considered variables related to school performance - class rank, teacher ratings, learning and behavior problems, dropout rates - as outcome variables [13, 6, 22–24]. Others have included individual achievement tests in addition to individual IQ tests. Bellinger, Stiles, and Needleman, for example, found that lead exposure at 2 years of age predicted academic deficits as measured by the Kaufman Test of Educational Achievement at 10 years of age . Lanphear et al. noticed an inverse relationship between lead exposure and the arithmetic and reading subtests of the Wide Range Achievement Test-Revised . Yule et al. found that scores on tests of reading, spelling, and IQ were negatively associated with blood lead levels . Fulton et al. noticed a dose-response relation between test scores on the British Abilities Scale and lead exposure . And Surkan et al. reported lower scores on the reading and math components on the Weschler Individual Achievement Test among children with even slightly elevated blood lead levels .
A study using school achievement test scores, carried out by Miranda et al. , assessed the relationship between early childhood lead exposure and school achievement, using the individual child as the unit of analysis. More than 8,000 children were in the sample. Their results indicated a dose-response effect between lead exposure and achievement on end-of-grade tests, for blood lead levels below 10, and even below 5, μg/dl. A subsequent study led by Miranda verified these conclusions with a much larger sample (over 57,000) of North Carolina school children, across all one hundred counties . A third study by Miranda et al. established a dose-response relationship for designations of exceptionality in the North Carolina dataset; the slope of the dose-response curve was even higher in the range between the first 1 and 5 μg/dl than between 5 and 10 .
In the current study, we used the results of regular academic testing in New York State, as well as separately published data on lead levels and income, to examine whether the academic performance of children in New York counties is associated with the incidence of elevated blood lead in those counties. Using publically accessible data, reported for the counties of New York, exclusive of New York City, we obtained statistics influenced by thousands of children. We included income data in the hope of controlling for socioeconomic status, addressing the question of how much the incidence of elevated blood lead predicts achievement, above and beyond the prediction accomplished by income.