Effects of measurement error on the strength of concentration-response relationships in aquatic toxicology

Derek L Sonderegger, Haonan Wang, Yao Huang, William H. Clements

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

The effect that measurement error of predictor variables has on regression inference is well known in the statistical literature. However, the influence of measurement error on the ability to quantify relationships between chemical stressors and biological responses has received little attention in ecotoxicology. We present a common data-collection scenario and demonstrate that the relationship between explanatory and response variables is consistently underestimated when measurement error is ignored. A straightforward extension of the regression calibration method is to use a nonparametric method to smooth the predictor variable with respect to another covariate (e.g., time) and using the smoothed predictor to estimate the response variable. We conducted a simulation study to compare the effectiveness of the proposed method to the naive analysis that ignores measurement error. We conclude that the method satisfactorily addresses the problem when measurement error is moderate to large, and does not result in a noticeable loss of power in the case where measurement error is absent.

Original languageEnglish (US)
Pages (from-to)824-828
Number of pages5
JournalEcotoxicology
Volume18
Issue number7
DOIs
StatePublished - Oct 2009
Externally publishedYes

Fingerprint

toxicology
Measurement errors
Toxicology
Ecotoxicology
ecotoxicology
Calibration
calibration
effect
method
simulation

Keywords

  • Arkansas River, CO
  • Ephemeroptera
  • Measurement error
  • Nonparametric function smoothing
  • Regression calibration

ASJC Scopus subject areas

  • Health, Toxicology and Mutagenesis
  • Toxicology
  • Management, Monitoring, Policy and Law

Cite this

Effects of measurement error on the strength of concentration-response relationships in aquatic toxicology. / Sonderegger, Derek L; Wang, Haonan; Huang, Yao; Clements, William H.

In: Ecotoxicology, Vol. 18, No. 7, 10.2009, p. 824-828.

Research output: Contribution to journalArticle

Sonderegger, Derek L ; Wang, Haonan ; Huang, Yao ; Clements, William H. / Effects of measurement error on the strength of concentration-response relationships in aquatic toxicology. In: Ecotoxicology. 2009 ; Vol. 18, No. 7. pp. 824-828.
@article{5c490c101ee345f78418a871c8e2e539,
title = "Effects of measurement error on the strength of concentration-response relationships in aquatic toxicology",
abstract = "The effect that measurement error of predictor variables has on regression inference is well known in the statistical literature. However, the influence of measurement error on the ability to quantify relationships between chemical stressors and biological responses has received little attention in ecotoxicology. We present a common data-collection scenario and demonstrate that the relationship between explanatory and response variables is consistently underestimated when measurement error is ignored. A straightforward extension of the regression calibration method is to use a nonparametric method to smooth the predictor variable with respect to another covariate (e.g., time) and using the smoothed predictor to estimate the response variable. We conducted a simulation study to compare the effectiveness of the proposed method to the naive analysis that ignores measurement error. We conclude that the method satisfactorily addresses the problem when measurement error is moderate to large, and does not result in a noticeable loss of power in the case where measurement error is absent.",
keywords = "Arkansas River, CO, Ephemeroptera, Measurement error, Nonparametric function smoothing, Regression calibration",
author = "Sonderegger, {Derek L} and Haonan Wang and Yao Huang and Clements, {William H.}",
year = "2009",
month = "10",
doi = "10.1007/s10646-009-0325-2",
language = "English (US)",
volume = "18",
pages = "824--828",
journal = "Ecotoxicology",
issn = "0963-9292",
publisher = "Springer Netherlands",
number = "7",

}

TY - JOUR

T1 - Effects of measurement error on the strength of concentration-response relationships in aquatic toxicology

AU - Sonderegger, Derek L

AU - Wang, Haonan

AU - Huang, Yao

AU - Clements, William H.

PY - 2009/10

Y1 - 2009/10

N2 - The effect that measurement error of predictor variables has on regression inference is well known in the statistical literature. However, the influence of measurement error on the ability to quantify relationships between chemical stressors and biological responses has received little attention in ecotoxicology. We present a common data-collection scenario and demonstrate that the relationship between explanatory and response variables is consistently underestimated when measurement error is ignored. A straightforward extension of the regression calibration method is to use a nonparametric method to smooth the predictor variable with respect to another covariate (e.g., time) and using the smoothed predictor to estimate the response variable. We conducted a simulation study to compare the effectiveness of the proposed method to the naive analysis that ignores measurement error. We conclude that the method satisfactorily addresses the problem when measurement error is moderate to large, and does not result in a noticeable loss of power in the case where measurement error is absent.

AB - The effect that measurement error of predictor variables has on regression inference is well known in the statistical literature. However, the influence of measurement error on the ability to quantify relationships between chemical stressors and biological responses has received little attention in ecotoxicology. We present a common data-collection scenario and demonstrate that the relationship between explanatory and response variables is consistently underestimated when measurement error is ignored. A straightforward extension of the regression calibration method is to use a nonparametric method to smooth the predictor variable with respect to another covariate (e.g., time) and using the smoothed predictor to estimate the response variable. We conducted a simulation study to compare the effectiveness of the proposed method to the naive analysis that ignores measurement error. We conclude that the method satisfactorily addresses the problem when measurement error is moderate to large, and does not result in a noticeable loss of power in the case where measurement error is absent.

KW - Arkansas River, CO

KW - Ephemeroptera

KW - Measurement error

KW - Nonparametric function smoothing

KW - Regression calibration

UR - http://www.scopus.com/inward/record.url?scp=69549091708&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=69549091708&partnerID=8YFLogxK

U2 - 10.1007/s10646-009-0325-2

DO - 10.1007/s10646-009-0325-2

M3 - Article

C2 - 19495966

AN - SCOPUS:69549091708

VL - 18

SP - 824

EP - 828

JO - Ecotoxicology

JF - Ecotoxicology

SN - 0963-9292

IS - 7

ER -