Alternatives For Handling Detection Limit Data in Impact Assessments

McBean, Edward A.; Rovers, Frank A.

doi:10.1111/j.1745-6592.1984.tb00881.x

Cited by 16 publications

(5 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Eliminating data or using zeros for results below detection limits introduces significant bias into statistical analyses (Fu & Wang, 2011; Helsel, 1990). Consistent with methods applied to groundwater samples (in which concentrations below detection limits can be frequent), results that were below detection limits were set to one half of the detection limit to reduce, although not eliminate, bias (Alley, 1990; McBean & Rovers, 1984). Critical r ‐values were determined from statistical tables on the basis of the number of matched samples in each dataset.…”

Section: Resultsmentioning

confidence: 99%

Bacterial and viral indicators of fecal contamination in drinking water

Plummer¹,

Long

Charest³

et al. 2014

Journal AWWA

View full text Add to dashboard Cite

A full report of this project, Assessment of Torque Teno Virus as a Viral Pathogen Indicator in Waters (4288), is available for free to Water Research Foundation sub scribers by logging on to www.waterrf.org.Plummer et al | http://dx.teaches classes on environmental engineering, laboratory methods, drinking water quality and treatment, and water quality modeling. Her research focuses on drinking water quality, source water protection, fecal pollution, microbial source tracking, indicator organisms, and alternative disinfection strategies. She holds MS and PhD degrees from

show abstract

Section: Resultsmentioning

confidence: 99%

Bacterial and viral indicators of fecal contamination in drinking water

Plummer¹,

Long

Charest³

et al. 2014

Journal AWWA

View full text Add to dashboard Cite

show abstract

“…Various investigators have addressed approaches for handling values below the detection limits, such as assigning these values as zero, one‐half the detection limit, or the detection limit [43]. For this risk assessment, copper and cadmium values below the detection limit were assumed to have log‐normal distributions.…”

Section: Exposure Characterizationmentioning

confidence: 99%

Ecological risk assessment of copper and cadmium in surface waters of Chesapeake Bay watershed

Hall

Scott

Killen

1998

Enviro Toxic and Chemistry

159

View full text Add to dashboard Cite

Abstract-This ecological risk assessment was designed to characterize risk of copper and cadmium exposure in the Chesapeake Bay watershed by comparing the probability distributions of environmental exposure concentrations with the probability distributions of species response data determined from laboratory studies. The overlap of these distributions was a measure of risk to aquatic life. Dissolved copper and cadmium exposure data were available from six primary data sources covering 102 stations in 18 basins in the Chesapeake Bay watershed from 1985 through 1996. Highest environmental concentrations of copper (based on 90th percentiles) were reported in the Chesapeake and Delaware (C and D) Canal, Choptank River, Middle River, and Potomac River; the lowest concentrations of copper were reported in the lower and middle mainstem Chesapeake Bay and Nanticoke River. Based on the calculation of 90th percentiles, cadmium concentrations were highest in the C and D Canal, Potomac River, Upper Chesapeake Bay, and West Chesapeake watershed. Lowest environmental concentrations of cadmium were reported in the lower and middle mainstem Chesapeake Bay and Susquehanna River. The ecological effects data used for this risk assessment were derived primarily from acute copper and cadmium laboratory toxicity tests conducted in both fresh water and salt water; chronic data were much more limited. The 10th percentile (concentration protecting 90% of the species) for all species derived from the freshwater acute copper toxicity database was 8

show abstract

“…Various strategies have been used to analyze the data with DLs, − such as deletion of the values below detection limits (BDLs) or simple substitution methods in which the measurements below detection limits are replaced with fixed values, such as zero, one-half of the DLs, or the DLs. , However, the deletion method may cause upward bias and a loss of information, and the substitution methods are generally not suitable when the results strongly depend on the substituted values . In particular, when there is a high proportion of censored data, the biased standard errors will further distort our inference. , If the distribution of the measurements is known to be either normal or log-normal, an alternative method is to fill in values randomly selected from the appropriate distribution or replace the values below the DLs with their conditional expected values (conditional on being less than the DLs). , This latter method produces unbiased regression parameter estimates and corrects for bias in variance estimates when the distributional assumptions are met, otherwise, bias will occur. ,− The robust method of regression on order statistics (ROS) suffers from the same problem ,,, ).…”

Section: Introductionmentioning

confidence: 99%

“…Various strategies have been used to analyze the data with DLs, [1][2][3][4][5][6][7][8][9][10][11][12] such as deletion of the values below detection limits (BDLs) or simple substitution methods in which the measurements below detection limits are replaced with fixed values, such as zero, one-half of the DLs, or the DLs. 1,2 However, the deletion method may cause upward bias and a loss of information, and the substitution methods are generally not suitable when the results strongly depend on the substituted values. 2 In particular, when there is a high proportion of censored data, the biased standard errors will further distort our inference.…”

Section: ' Introductionmentioning

confidence: 99%

Nonparametric Rank Regression for Analyzing Water Quality Concentration Data with Multiple Detection Limits

Wang

2011

Environ. Sci. Technol.

View full text Add to dashboard Cite

Environmental data usually include measurements, such as water quality data, which fall below detection limits, because of limitations of the instruments or of certain analytical methods used. The fact that some responses are not detected needs to be properly taken into account in statistical analysis of such data. However, it is well-known that it is challenging to analyze a data set with detection limits, and we often have to rely on the traditional parametric methods or simple imputation methods. Distributional assumptions can lead to biased inference and justification of distributions is often not possible when the data are correlated and there is a large proportion of data below detection limits. The extent of bias is usually unknown. To draw valid conclusions and hence provide useful advice for environmental management authorities, it is essential to develop and apply an appropriate statistical methodology. This paper proposes rank-based procedures for analyzing non-normally distributed data collected at different sites over a period of time in the presence of multiple detection limits. To take account of temporal correlations within each site, we propose an optimal linear combination of estimating functions and apply the induced smoothing method to reduce the computational burden. Finally, we apply the proposed method to the water quality data collected at Susquehanna River Basin in United States of America, which clearly demonstrates the advantages of the rank regression models.

show abstract

Alternatives For Handling Detection Limit Data in Impact Assessments

Cited by 16 publications

References 1 publication

Bacterial and viral indicators of fecal contamination in drinking water

Bacterial and viral indicators of fecal contamination in drinking water

Ecological risk assessment of copper and cadmium in surface waters of Chesapeake Bay watershed

Nonparametric Rank Regression for Analyzing Water Quality Concentration Data with Multiple Detection Limits

Contact Info

Product

Resources

About