On the first-order bias of parameter estimates in a quantal response model under alternative estimation procedures

Abstract

SUMMARY To estimate the parameters in a quantal response model, various alternative procedures have been used. A comparison is made between maximum likelihood, minimum chisquared and modified minimum chi-squared in terms of first-order bias of parameter estimates, with particular reference to the standard probit model. It is shown that, in this respect, the procedures may differ widely, with maximum likelihood seeming the most satisfactory of the three alternatives.