Thomas Lumley • 1/26/2019

Two quick survey items

This technical article discusses statistical sampling methods for case-control studies, specifically focusing on optimal stratified sampling (Neyman allocation) for estimating logistic regression coefficients. It analyzes the influence functions and variances for cases and controls, concluding that similar sample sizes are optimal for small covariate effects. The article also explains variance estimation in two-phase sampling designs, referencing implementation details in the R 'survey' package.

0 comments

#Logistic Regression #Statistical Sampling #Influence Functions