Nondetects and data analysis : statistics for censored environmental data

Title:

Personal Author:

Helsel, Dennis R.

Publication Information:

Hoboken, NJ : Wiley-Interscience, 2005

ISBN:

9780471671732

Subject Term:

Environmental sciences -- Statistical methods

Survival analysis (Biometry)

Available:*

Library	Item Barcode	Call Number	Material Type	Item Category 1	Status
Searching... PSZ JB	30000010077928	QE45.S73 H44 2005	Unknown	1:CHECKING	Searching... Unknown

Summary

STATISTICS IN PRACTICE

Statistical methods for interpreting and analyzing censored environmental data

Nondetects And Data Analysis: Statistics for Censored Environmental Data provides solutions for environmental scientists and professionals who need to interpret and analyze data that fall below the laboratory detection limit. Adapting survival analysis methods that have been successfully used in medical and industrial research, the author demonstrates, for the first time, their practical applications for studies of trace chemicals in air, water, soils, and biota. Readers quickly become proficient in these methods through the use of real-world examples that are solved using MINITAB® Release 14, a popular statistical software package, as well as other commonly used software packages.

Everything needed to master these innovative statistical methods is provided, including:

Accompanying Web site featuring answers to book exercises and datasets, as well as MINITAB® macros to perform methods, which are not available in the commercial version Methods for data with multiple detection limits Solutions for research studies in which all data are below detection limits Techniques for constructing confidence, prediction, and tolerance intervals for data with nond-tects Methods for data with multiple detection limits

Chapters are organized by objective, such as computing intervals, comparing groups, and correlations, which enables readers to more easily apply the text to their particular research and goals. Extensive references to the literature for more in-depth research are provided; however, the text itself avoids complex math and calculus making it accessible to anyone in the environmental sciences. Environmental scientists and professionals will find the hands-on guidance and practical examples invaluable.

Author Notes

DENNIS R. HELSEL, PhD, is a 2003 recipient of the American Statistical Association Section on Statistics and the Environment's Distinguished Achievement Award. His textbook, Statistical Methods in Water Resources, coauthored by Robert Hirsch, is a standard reference in the water resources community. Dr. Helsel currently does research on behalf of the United States Geological Survey in Denver, Colorado.

Preface	p. xiii
Acknowledgments	p. xv
Introduction: An Accident Waiting to Happen	p. 1
1 In Focus	p. 5
2 The Three Approaches	p. 9
Approach 1 Substitution	p. 12
Approach 2 Maximum likelihood estimation	p. 13
Approach 3 Nonparametric methods	p. 15
Application of survival analysis methods to environmental data	p. 17
Power transformations with survival analysis	p. 18
Parallels to uncensored methods	p. 19
3 Reporting Limits	p. 21
Limits when the standard deviation is considered constant	p. 21
The detection limit	p. 22
The quantitation limit	p. 25
False negatives	p. 27
Data between the limits	p. 28
Informative censoring - biasing interpretations	p. 30
Limits when the standard deviation changes with concentration	p. 33
For further study	p. 34
4 Storing Data in Databases	p. 37
Method 1 Negative numbers	p. 37
Method 2 Indicator variable	p. 38
Method 3 Interval endpoints	p. 39
5 Plotting Data with Nondetects	p. 43
Boxplots	p. 43
Histogram	p. 45
Probability plot	p. 45
Empirical distribution function	p. 47
Survival function plots	p. 49
X-Y scatterplots	p. 50
6 Computing Summary Statistics	p. 55
Substitution (fabrication) methods	p. 55
Maximum likelihood estimation	p. 56
Nonparametric method: Kaplan-Meier	p. 63
ROS: A "robust" method	p. 68
A review of comparison studies	p. 73
7 Computing Interval Estimates	p. 81
Parametric intervals	p. 82
Nonparametric intervals	p. 84
Intervals for censored data by substitution	p. 85
Intervals for censored data by maximum likelihood	p. 86
Intervals using 'robust' parametric methods	p. 103
Nonparametric intervals for censored data	p. 103
Bootstrapped intervals	p. 113
8 What Can Be Done When All Data Are Below the Reporting Limit?	p. 119
Point estimates	p. 119
Probability of exceeding the reporting limit	p. 121
Exceedance probability for a standard higher than the reporting limit	p. 125
Hypothesis tests between groups	p. 128
Summary	p. 129
9 Comparing Two Groups	p. 131
Substitution methods	p. 132
Maximum likelihood estimation	p. 134
Nonparametric methods	p. 139
Value of the information in nondetects	p. 151
Transformations with score tests	p. 151
Paired observations	p. 153
Comparing data to a standard using paired tests	p. 161
Summary of two-sample tests for censored data	p. 162
10 Comparing Three or More Groups	p. 165
Substitution methods	p. 165
Maximum likelihood estimation	p. 167
Nonparametric methods	p. 176
Summary	p. 183
11 Correlation	p. 185
Types of correlation coefficients	p. 185
Substitution methods	p. 186
Maximum likelihood estimation	p. 187
Nonparametric methods	p. 188
Summary. A comparison among methods	p. 194
For further study	p. 196
12 Regression and Trends	p. 199
Substitution methods	p. 200
Maximum likelihood estimation	p. 201
Theil-Sen nonparametric regression	p. 209
Logistic regression	p. 215
Additional methods for censored regression	p. 224
Appendix Datasets	p. 229
References	p. 237
Index	p. 247

Available:*

On Order

Summary

Summary

Author Notes

Table of Contents