The purpose of this document is to report the results of statistical analysis of this extensive residential tapwater lead database. Statistical analyses were chosen within the context of providing EPA and water utilities with information relevant to residential lead contamination dynamics and potential technical and regulatory measures for effective control. The database has the important advantage of containing a large number of residential samples representing a wide range of individual and public water sources, pH levels, corrosion control additives. water hardness, and plumbing system ages. Relatively rigorous statistical analyses were performed and are presented with only minimal interpretation and discussion.

The following are some of the major findings which are statistically documented in the body of this report.

- Approximately 25.8% of the 2724 residents included in this study have a 500 ml first draw lead concentration of 10 parts per billion (ppb) or greater. The corresponding percent for one-liter first draw (1LFD) samples is 17.9%.

- The 80th, 90th, and 95th concentration percentiles for all 1LFD samples are 9.23, 18.4, and 30.0 ppb. respectively.

- Only 2.6% of all residences have at-the-tap lead levels above 10 ppb after running the tap at full volume for one minute. About 92% of the residences with 1LFD levels greater than 10 ppb will drop below 10 ppb with a one minute purge.

- NC residences on private well or spring water have a significantly greater risk of tapwater lead contamination than residences on municipal supplies.

- About 16% of NC residences on municipal supplies have 1LFD lead levels of 10 ppb or greater with 80th, 90th, and 9.5th concentration percentiles of 7.91, 16.9, and 29.4 respectively. These results suggest that a majority of NC water utilities would not meet a criterion of a 90th percentile concentration of 15 ppb.

- For reasons that are unclear, residences with PVC or other plastic plumbing systems have essentially the same risk of lead contamination as residences with copper-solder systems. A separate study of faucet fixtures indicates that this phenomenon can not be explained by meal faucet fixture lead contributions alone .

- Residences and plumbing systems between three and nine years of age have the greatest risk of lead contaminalion, while those over 30 years old have the least risk. However even for the 30+ age group, 9.4% of the residences on municipal water supplies have 1LFD levels of 10 ppb or greater.

- A number of different analyses provide strong, albeit indirect, evidence of substantial non-compliance with the December, 1986 regulations banning the use of leaded solder and pipe.

- Various corrosion inhibitors appear to significantly reduce the prevalence and level of residential lead contamination. The effectiveness of some types, particularly the metaphosphates, is very dependent on water hardness, being much less effective for very soft waters.

- Increased pH is significantly correlated with decreased lead concentrations. However, although statistically significant, the effect is actually very small.

- Water hardness and residential lead concentrations are more strongly correlated than pH and lead concentrations. Again, however, the total percentage of lead concentration variability explained by differences in water hardness is small.

- A regression model incorporating pH, hardness, corrosion inhibitors, and plumbing system age, as well as the interactive effects of these variables explains only 15% of the observed lead concentration variability in municipally-supplied residences. Clearly, the primary determinate of residential lead levels appears to be the amount or surface area of lead exposed within the plumbing system.