|
|
|
The cover illustrates performance test results for a particular commercial spot-test kit for lead in paint. In a statistically-designed experiment, several different test kits were compared on samples with different levels of lead, for combinations of various other factors. The curves are from a logistic model with operator (5 levels), lead type (white or yellow), substrate (reactive or nonreactive) and lead concentration (10 levels) as explanatory variables. Colored curves indicate individual operators. Proportions of positive test results are displayed as points with binomial 95% confidence intervals, where appropriate. The shaded areas indicate lead levels below the regulatory limit for remediation. The broken curves are from a previously-published EPA field study. Both lead type and substrate type can be seen to be important, and there also appears to be considerable variability in the performance of the operators. 
Date created: 7/20/2001 |