| Abstract: | This paper examines the use of order statistics to provide distribution-free point and confidence interval estimates of quantiles of the distribution of contaminants in environmental media. These procedures are straightforward to implement and are often unaffected by the presence of observations below the detection limit. This approach to the estimation of quantiles has been implemented in the determination of the Ontario typical range of chemical parameters in soil, vegetation, moss bags and snow. These distributions range from symmetric to highly skewed to the right, with zero to 100% of observations below the detection limit. |
