More than obvious: Better methods for interpreting nondetect data

Fifteen years after writing a feature article about methods for the statistical analysis of non-detect data-those concentrations that are only known to be somewhere between zero and a laboratory's reporting level-Dennis R. Helsel of the U.S. Geological Survey returns to survey the current state of the art. Although the statistical methods have improved, regulatory guidance remains mired in old and sometimes untrustworthy methods. © 2005 American Chemical Society.