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

Statistical evidence for GLM regression parameters: a robust likelihood approach.

Jeffrey D Blume1, Li Su, Remigio M Olveda

  • 1Center for Statistical Sciences, Brown University, Providence RI 02912, USA. jblume@stat.brown.edu

Statistics in Medicine
|January 11, 2007
PubMed
Summary

Statistical model misspecification can yield misleading evidence. This study introduces a robust adjusted likelihood method for generalized linear models, ensuring reliable statistical evidence even with imperfect models.

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

The Impact of the 2023 American Cancer Society Screening Recommendations on Racial, Ethnic, and Sex Disparities in Lung Cancer Screening Eligibility.

Chest·2026
Same author

Addressing algorithmic bias in lung cancer screening eligibility.

Journal of the National Cancer Institute·2025
Same author

Evaluating the Performance and Clinical Utility of AI-driven Diagnostic Tools in Radiology.

Radiology·2025
Same author

Individual- and Group-Level Disparities Between Racial and Ethnic Groups in Lung Cancer Screening Eligibility Criteria.

JAMA network open·2025
Same author

Sequential monitoring using the Second Generation P-Value with Type I error controlled by monitoring frequency.

The American statistician·2025
Same author

The Thoracic Research Evaluation and Treatment 2.0 Model: A Lung Cancer Prediction Model for Indeterminate Nodules Referred for Specialist Evaluation.

Chest·2023

Area of Science:

  • Statistics
  • Biostatistics
  • Statistical Modeling

Background:

  • The reliability of likelihood ratios for hypothesis testing depends heavily on the accuracy of the working model.
  • Misspecified models can lead to unreliable or misleading statistical evidence.
  • Existing methods by Royall and Tsou offer robustness but require adaptation for specific regression settings.

Purpose of the Study:

  • To extend the robust adjusted likelihood methodology to the generalized linear model (GLM) regression framework.
  • To enhance the reliability of statistical evidence measurement in regression analysis.
  • To make likelihood-based evidence assessment more accessible and viable for complex data.

Main Methods:

  • Application and extension of the robust adjusted likelihood concept to GLM regression.

Related Experiment Videos

  • Development of adjustment factors for misspecified working models within GLMs.
  • Utilizing simulated and real-world data (parasitic infection rates) for illustration and validation.

    • Main Results:

    • The robust adjusted likelihood provides asymptotically reliable statistical evidence, comparable to correctly specified models.
    • Adjustment factors for GLMs can be derived using standard statistical software.
    • The approach shows connections to the sandwich estimator for robust standard errors in regression.

    Conclusions:

    • The robust adjusted likelihood method effectively addresses model misspecification in GLM regression.
    • This work broadens the applicability of likelihood methods for measuring statistical evidence in regression settings.
    • The findings support the use of adjusted likelihoods for more dependable statistical inference.