Effects of Measurement Errors on Estimates of Exposure-Response Relationships

  • B. Armstrong
Part of the Recent Results in Cancer Research book series (RECENTCANCER, volume 120)


Establishing a causal association between an occupational exposure and the incidence of cancer requires the consideration of many criteria. One of these is whether an association between level of exposure and risk has been demonstrated. For this purpose an association between an ordinal measure of exposure (e.g. three exposure categories) and risk is generally considered sufficient. Once a causal association has been established with reasonable certainty, interest often focusses on a more quantitative relationship between an absolute measure of exposure and risk — for a given quantity of exposure, what is the increase in risk? In particular, estimates of such relationships are required to inform the process of setting standards for “acceptable” limits of exposure. This paper is relevant mainly to these quantitative estimates of exposure-response relationships.


Standardize Mortality Ratio Measurement Error Model Measurement Error Variance True Exposure Covariate Measurement Error 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. Armstrong BG (1985) Measurement error in the generalised linear model. Comm Statist Simul Comp 14: 529–544CrossRefGoogle Scholar
  2. Armstrong BG, Oakes D (1982) The effects of approximation in exposure assessments on estimates of exposure response relationships. Scand J Work Environ Health 8 [Suppl 1]: 20–23PubMedCrossRefGoogle Scholar
  3. Armstrong BG, Tremblay CG, Cyr D, Thériault GP (1986) Estimating the relationship between exposure to tar volatiles and the incidence of bladder cancer in aluminum smelter workers. Scand J Work Environ Health 12: 486–493PubMedCrossRefGoogle Scholar
  4. Armstrong BG, Whittemore AS, Howe GR (1989) Analysis of case-control data with covariate measurement error: application to diet and colon cancer. Stat Med 8: 1151–1165PubMedCrossRefGoogle Scholar
  5. Berkson J (1950) Are there two regressions? J Am Stat Assoc 45: 164–180CrossRefGoogle Scholar
  6. Carroll RJ, Spiegelman CH, Lan KK, Bailey KT, Abbott RD (1984) On errors-invariables for binary regression models. Biometrika 71: 19–25CrossRefGoogle Scholar
  7. Chen TT (1989) Overview of misclassification in epidemiology. Stat Med 8: 1095–1106PubMedCrossRefGoogle Scholar
  8. Clayton DG (1988) Models for the analysis of cohort and case-control studies with inaccurately measured exposures. In: Dwyer JH, Lippert P, Feinleib M, Hoffmeister H (eds) Statistical models for longitudinal studies of health. Oxford University Press, New YorkGoogle Scholar
  9. Cochran WG (1968) Errors of measurement in statistics. Technometrics 10: 637–666CrossRefGoogle Scholar
  10. Doll R, Peto R (1978) Cigarette smoking and bronchial carcinoma: dose and time relationships among regular and lifelong nonsmokers. J Epidemiol Community Health 32: 303–313CrossRefGoogle Scholar
  11. Fuller WA (1987) Measurement error models. Wiley, New YorkCrossRefGoogle Scholar
  12. Goldberg MS, Siemiatycki J, Gerin M (1986) Inter-rater agreement in assessing occupational exposure in a case-control study. Br J Ind Med 43: 667–676PubMedGoogle Scholar
  13. Greenland S (1980) The effect of misclassification in the presence of covariates. Am J Epidemiol 112: 564–569PubMedGoogle Scholar
  14. Hanley J, Liddell FDK (1985) Fitting relationships between exposure and standardized mortality ratios. J Occup Med 27: 555–560PubMedCrossRefGoogle Scholar
  15. Kelsey JL, Thompson WD, Evans AS (1986) Methods in observational epidemiology. Oxford University Press, New York, pp 285–308Google Scholar
  16. Kendall M, Stuart A (1979) The advanced theory of statistics vol 2. MacMillan, New York, pp 399–443Google Scholar
  17. Kupper L (-1984) Effects of the use of unreliable surrogate variables on the validity of epidemiological research studies. Am J Epidemiol 120: 643–648PubMedGoogle Scholar
  18. Lagakos S (1987) Effects of mismodelling and mismeasuring explanatory variables on tests of their association with a response variable. Stat Med 7: 257–274CrossRefGoogle Scholar
  19. McDonald JC, McDonald AD, Armstrong BG, Sebastien P (1986) Cohort mortality study of vermiculite miners exposed to tremolite. Br J Ind Med 43: 436–444PubMedGoogle Scholar
  20. Pepe M, Self SG, Prentice RL (1989) Further results on covariate measurement errors in cohort studies with time to response data. Stat Med 8: 1167–1178PubMedCrossRefGoogle Scholar
  21. Prentice RL (1982) Covariate measurement error and parameter estimation in a failure time regression model. Biometrika 69: 331–342CrossRefGoogle Scholar
  22. Prentice RL, Farewell VT (1986) Relative risk and odds ratio regression. Annu Rev Public Health 7: 35–58PubMedCrossRefGoogle Scholar
  23. Schafer DW (1987) Covariate measurement error in generalised linear models. Biometrika 74: 385–391CrossRefGoogle Scholar
  24. Schafer DW, Stefanski LA, Tosteson TD (1989) A measurement error model for binary and ordinal regression. Stat Med 8: 1139–1138PubMedCrossRefGoogle Scholar
  25. Snedecor GW, Cochran WG (1967) Statistical methods. Iowa State University Press, Iowa, pp 164–167Google Scholar
  26. Stefanski LA, Carroll RJ (1985) Covariate measurement error in logistic regression. Ann Stat 13: 1335–1351CrossRefGoogle Scholar
  27. Thériault G, Tremblay C, Cordier S, Gingras S (1984) Bladder cancer in the aluminium industry. Lancet is 947–950Google Scholar
  28. Tosteson TD, Tsiatis AA (1988) The asymptotic relative efficiency of score tests in the generalized linear model with surrogate covariates. Biometrika 75: 507–514CrossRefGoogle Scholar
  29. Whittemore AS, Grosser S (1986) Regression methods for data with incomplete covariates. In: Moolgavkar SH, Prentice RR (eds) Modern statistical methods in chronic disease epidemiology. Wiley, New York, pp 19–34Google Scholar
  30. Willett W (1989) An epidemiologic perspective on exposure measurement error. Stat Med 8: 1031–1040PubMedCrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin·Heidelberg 1990

Authors and Affiliations

  • B. Armstrong
    • 1
  1. 1.School of Occupational HealthMcGill UniversityMontrealCanada

Personalised recommendations