A note on estimation for quantal response data

Abstract

Expressions are presented which show why minimum chi-square estimation for quantal response data consistently gives smaller slope estimates than maximum-likelihood estimation, and why the corresponding LD 50 estimates display no such consistent relationship.