Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Blind assignment of exposure does not always prevent differential misclassification.

S Wacholder1, M Dosemeci, J H Lubin

  • 1Epidemiology and Biostatistics Program, National Cancer Institute, Bethesda, MD 20892.

American Journal of Epidemiology
|August 15, 1991
PubMed
Summary

Exposure measurement error can lead to differential misclassification in epidemiological studies. Collapsing exposure categories can introduce bias, even if initial misclassification is nondifferential.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Investing in prospective cohorts for etiologic study of occupational exposures.

American journal of industrial medicine·2015
Same author

Plasma cytokine levels and human papillomavirus infection at the cervix in rural Nigerian women.

Cytokine·2013
Same author

Misunderstandings in the misconception on the use of pack-years in analysis of smoking.

British journal of cancer·2013
Same author

Risk of renal cell carcinoma in relation to blood telomere length in a population-based case-control study.

British journal of cancer·2011
Same author

Meat-cooking mutagens and risk of renal cell carcinoma.

British journal of cancer·2011
Same author

The exposure metric: does including time since exposure in the calculation of working lifetime exposure provide a better understanding of disease risk than the cumulative exposure?

La Medicina del lavoro·2011

Area of Science:

  • Epidemiology
  • Biostatistics

Background:

  • Exposure measurement error is a common challenge in epidemiological research.
  • Differential misclassification can significantly bias study results.
  • Nondifferential misclassification is often assumed to bias results towards the null.

Purpose of the Study:

  • To investigate the potential for differential misclassification arising from collapsing exposure categories.
  • To demonstrate how nondifferential misclassification in a larger table can lead to differential misclassification in a collapsed table.
  • To explain the mechanisms by which collapsing exposure levels can induce bias.

Main Methods:

  • Presentation of hypothetical examples using 2x3 and 2x2 contingency tables.
  • Analysis of odds ratios under different exposure misclassification scenarios.

Related Experiment Videos

  • Explanation of bias induction through category collapsing.

    • Main Results:

    • Collapsing exposure categories can induce differential misclassification, even when initial misclassification is nondifferential.
    • Observed odds ratios in collapsed tables can be in the opposite direction or more extreme than true values.
    • Bias can occur regardless of whether collapsing happens during analysis or exposure assessment.

    Conclusions:

    • Certainty regarding nondifferential misclassification is not guaranteed, even with careful exposure assessment.
    • Collapsing exposure categories is a critical step that requires careful consideration to avoid introducing bias.
    • Blind exposure assessment does not prevent the induction of differential misclassification through category collapsing.