Simulation study shows g-computation for risk differences often inflates Type I error in very small samples under standard Wald inference, while robust variants and classical Mantel-Haenszel methods control errors better.
On Generalized Score Tests
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ME 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Assessing covariate-adjusted risk differences in small-sample clinical trials
Simulation study shows g-computation for risk differences often inflates Type I error in very small samples under standard Wald inference, while robust variants and classical Mantel-Haenszel methods control errors better.