Heteroskedasticity variance estimators exhibit a predictable direction of bias
the verdict
SUPPORTED
the evidence backs this
confidence 79/100
Heteroskedasticity and clustered variance estimators frequently exhibit predictable directions of bias, such as underestimating standard errors in specific small-sample or non-standard settings.
Evidence for · 2
A Comparison of Variance Estimators for Logistic Regression Models Estimated Using Generalized Estimating Equations (GEE) in the Context of Observational Health Services Research.
2024 · cited by 3
Paper [6] notes that conventional variance estimators, such as the Liang-Zeger estimator, systematically underestimate standard errors (exhibiting a predictable downward bias) when cluster counts are small.
See more details
More for · 1
A practice-oriented guide to statistical inference in linear modeling for non-normal or heteroskedastic error distributions.
2025 · cited by 1
Paper [10] evaluates standard error bias under heteroskedasticity and finds that classical estimators and specific HC alternatives display predictable patterns of bias depending on the scenario and sample size.