Mathematics
8 years ago

Misspecified autocorrelation structures in data analysis lead to inaccurate parameter estimates.

International Journal of Statistics in Medical Research
Specification of Variance-Covariance Structure in Bivariate Mixed Model for Unequally Time-Spaced Longitudinal Data
Melike Bahçecitapar, Özge Karadağ, Serpil Aktaş

Paper Summary

Related papers

Theme background image
Mathematics
8 years ago

New Stata Module Revolutionizes Data Visualization with Combineplot Feature

Statistical Modeling
Sensory Analysis
Statistical Literacy
1 view
Theme background image
Computer Science
3 years ago

New study reveals how two factors impact frequency of observations.

Inductive Modeling
Multicollinearity
Model-Based Clustering
2 views
8 years ago

New study reveals simple way to choose correct bivariate statistic!

Researchers investigated using multivariate models to analyze longitudinal data sets with multiple response variables and covariates in medical studies. By examining correlations between response variables and autocorrelation over time, they found that accurately specifying the autocorrelation structure is crucial. Comparing simpler models with multivariate models, they discovered that incorrect specification of autocorrelation negatively impacts parameter estimates.