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

Incorporating missingness for estimation of marginal regression models with multiple source predictors.

Heather J Litman1, Nicholas J Horton, Bernardo Hernández

  • 1New England Research Institutes, 9 Galen St, Watertown, MA 02472, USA. hlitman@neriscience.com

Statistics in Medicine
|June 7, 2006
PubMed
Summary

This study introduces a new Maximum Likelihood (ML) method for analyzing multiple informant data, even with missing information. The ML approach is more efficient and robust than existing Generalized Estimating Equations (GEE) methods.

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

Explainable machine learning in healthcare: methods, interpretation, and applications for clinical research.

Journal of the American Medical Informatics Association : JAMIA·2026
Same author

Effectiveness and persistence of biological and targeted synthetic treatment in patients with rheumatoid arthritis and chronic kidney disease: a multicentre, prospective cohort study.

The Lancet. Rheumatology·2026
Same author

Features Associated with Therapy Switch Among PPD CorEvitas Psoriasis Registry Patients.

Dermatology and therapy·2026
Same author

Development and External Validation of a Multivariable Predictive Model for Progression to Difficult-to-Treat Rheumatoid Arthritis in Biologic-Experienced Patients.

Arthritis care & research·2025
Same author

A results to action framework for community verification: A case study from a performance based financing program in Zimbabwe.

PLOS global public health·2025
Same author

Comprehensive Measurements and Analyses of Ear Canal Geometry From Late Infancy Through Late Adulthood: Age-Related Variations and Implications for Basic Science and Audiological Measurements.

Trends in hearing·2025

Area of Science:

  • Psychometrics
  • Statistical modeling
  • Behavioral science

Background:

  • Multiple informant data is crucial for measuring constructs accurately.
  • Incomplete observations are common in multi-informant studies, posing analytical challenges.
  • Existing methods like Generalized Estimating Equations (GEE) may not fully address substantial missingness.

Purpose of the Study:

  • To introduce a novel Maximum Likelihood (ML) technique for analyzing multiple informant data with missing predictors and responses.
  • To compare the performance of the ML technique against the Generalized Estimating Equations (GEE) approach.
  • To assess the robustness and efficiency of the ML method under various missing data conditions.

Main Methods:

  • Development of a Maximum Likelihood (ML) estimation technique for multi-informant models.

Related Experiment Videos

  • Analytical comparison of ML with Generalized Estimating Equations (GEE).
  • Simulation studies to evaluate ML performance with missing data (MCAR) and departures from normality.
  • Application of both ML and GEE to real-world data on physical activity and obesity.

    • Main Results:

    • The ML technique effectively handles missing data in multi-informant models.
    • ML demonstrates greater efficiency compared to GEE in simulations with missing data.
    • The ML approach is robust to deviations from normality under Missing Completely At Random (MCAR) conditions.
    • ML allows for direct comparison of informant effects without data standardization.

    Conclusions:

    • The proposed ML method offers a more efficient and robust alternative for analyzing multiple informant data, especially when missingness is present.
    • This technique provides a valuable tool for researchers dealing with incomplete multi-informant datasets.
    • The findings support the utility of ML in understanding complex relationships using diverse data sources.