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

Sensitivity analysis after multiple imputation under missing at random: a weighting approach.

James R Carpenter1, Michael G Kenward, Ian R White

  • 1Medical Statistics Unit, London School of Hygiene and Tropical Medicine, Keppel Street, London, UK. james.carpenter@lshtm.ac.uk

Statistical Methods in Medical Research
|July 11, 2007
PubMed
Summary
This summary is machine-generated.

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

European Society of Contact Dermatitis Guideline for Diagnostic Patch Testing-Recommendations on Best Practice (Update 2026).

Contact dermatitis·2026
Same author

SPIRIT 2025 statement: updated guideline for protocols of randomised trials.

Lancet (London, England)·2026
Same author

Sample Size Calculation for the ROCI Design.

Statistics in medicine·2026
Same author

Bayesian analysis in confirmatory clinical trials: A narrative review and discussion of current practice.

Clinical trials (London, England)·2026
Same author

Adjusting for confounding in population administrative data when confounders are only measured in a linked cohort.

International journal of population data science·2026
Same author

An investigation of the impact of using contrast- and arm-synthesis models for network meta-analysis.

Research synthesis methods·2026
Same journal

A joint model for a longitudinal outcome and a progressive multistate model under a mixed observation scheme.

Statistical methods in medical research·2026
Same journal

Efficient semi-supervised estimation of optimal individualized treatment regimes with survival outcome.

Statistical methods in medical research·2026
Same journal

Asymptotic online FWER control for dependent test statistics.

Statistical methods in medical research·2026
Same journal

Regression analysis of misclassified current status data with potentially unknown test accuracy.

Statistical methods in medical research·2026
Same journal

Bayesian multivariate linear mixed-effects models with varied association structures.

Statistical methods in medical research·2026
Same journal

Inference about the ratio of age-standardized rates between two overlapping populations.

Statistical methods in medical research·2026
See all related articles

This study introduces a novel method for handling missing data in statistical analyses. It offers a practical approach to assess the impact of data not missing at random (NMAR) on multiple imputation results.

Area of Science:

  • Statistics
  • Biostatistics
  • Data Science

Background:

  • Multiple imputation (MI) is a standard method for analyzing data with missing values.
  • Current MI methods often assume data are missing at random (MAR), which may not hold true in real-world scenarios.
  • Assessing sensitivity to the MAR assumption and analyzing data not missing at random (NMAR) remains challenging due to limited software and complex modeling requirements.

Purpose of the Study:

  • To propose a simple and practical alternative method for analyzing data with missing values, particularly when the missingness mechanism might be not at random (NMAR).
  • To provide a way to assess the sensitivity of multiple imputation analyses to the MAR assumption.
  • To offer a method that can approximate or check the results of full NMAR modeling.

Main Methods:

Related Experiment Videos

  • The proposed method involves performing multiple imputation under the MAR assumption.
  • Parameter estimates are obtained for each imputed dataset.
  • A weighted average of these estimates is calculated, with weights reflecting the degree of departure from MAR, to produce an overall NMAR parameter estimate.

Main Results:

  • The proposed weighted averaging approach can yield results comparable to joint modeling in some situations as the number of imputations increases.
  • In other cases, it provides approximate estimates for full NMAR modeling, indicating the necessity and validating results.
  • The method was illustrated through a simulation study and applied to real-world data from a peer review quality improvement trial.

Conclusions:

  • The developed method offers a valuable tool for sensitivity analyses in multiple imputation, especially when NMAR data is suspected.
  • It provides a computationally feasible alternative to complex joint modeling for NMAR data.
  • This approach enhances the robustness and interpretability of statistical analyses involving missing data.