Arugaslan, Cook, and Kieschnick (2004) challenge underpricing results obtained from conventional cross-sectional regression analysis on the grounds that standard methods fail to properly account for underwriter price stabilization and adequately capture variations in information asymmetries related to firm size. We find that results from the long-standing methods for estimating underpricing relations are robust to one's choice of size proxy. We obtain consistent estimates of underpricing determinants from censored regressions of first-day returns and from least squares regressions of longer horizon returns, whereas estimates from the mixed distribution proposed by Arugaslan et al. are sensitive to distribution assumptions and starting values.