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

Bayesian Sensitivity Analysis of a Nonlinear Dynamic Factor Analysis Model with Nonparametric Prior and Possible

Niansheng Tang1, Sy-Miin Chow2, Joseph G Ibrahim3

  • 1Department of Statistics, Yunnan University, Kunming, People's Republic of China.

Psychometrika
|October 15, 2017
PubMed
Summary

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

Statistics and AI - A Fireside Conversation.

Harvard data science review·2026
Same author

Cardiovascular-Kidney-Metabolic Syndrome: Conceptualising an Approach to Health Economic Modelling.

Diabetes, obesity & metabolism·2026
Same author

Artificial Intelligence in Image-Based Cardiovascular Disease Analysis.

Annual review of biomedical data science·2026
Same author

Multi-organ imaging and genetics show the impact of sleep patterns on the human brain and body.

Communications medicine·2026
Same author

Scalable subclonal reconstruction of cancer cells in DNA sequencing data using a penalized likelihood model.

bioRxiv : the preprint server for biology·2026
Same author

Connectome-based spatial statistics enabling large-scale population analyses of human connectome across cohorts.

bioRxiv : the preprint server for biology·2026

This study introduces a Bayesian local influence method for sensitivity analysis in dynamic factor analysis models. It efficiently detects outlying cases and model misspecifications, improving the reliability of psychological research using complex data.

Area of Science:

  • Psychometrics
  • Statistical Modeling
  • Time Series Analysis

Background:

  • Psychological concepts are often unobserved latent factors inferred from multiple indicators.
  • Dynamic factor analysis (DFA) models with random effects analyze within- and between-person variations in multivariate time series data.
  • Dirichlet process (DP) priors allow flexibility in modeling individual-specific time series parameters beyond standard distributional assumptions.

Purpose of the Study:

  • To develop a computationally feasible Bayesian local influence method for comprehensive sensitivity analysis in complex DFA models.
  • To enable simultaneous assessment of multiple modeling components within a single model fitting procedure.
  • To identify sensitive modeling components and potential sources of misspecification.

Main Methods:

Keywords:
Bayesian local influenceBayesian perturbation manifoldDirichlet process priornonignorable missing datanonlinear dynamic factor analysis modelsensitivity analysis

Related Experiment Videos

  • Proposed a Bayesian local influence approach for sensitivity analysis.
  • Applied the method to dynamic factor analysis models incorporating random effects and Dirichlet process priors.
  • Utilized five illustrations and one empirical example for validation.

Main Results:

  • The proposed method effectively detects outlying cases and common sources of model misspecification.
  • It facilitates the identification of modeling components sensitive to changes in the Dirichlet process prior specification.
  • Demonstrated the utility of the approach in enhancing the robustness of DFA models.

Conclusions:

  • The Bayesian local influence method offers a critical and computationally efficient tool for sensitivity analysis in complex statistical models.
  • This approach improves the reliability and interpretability of findings from dynamic factor analysis, particularly in psychological research.
  • It aids researchers in understanding model behavior and identifying areas requiring careful specification and validation.