Schmidt, Hunter, McKenzie and Muldrow's (1979) global estimation procedure for determining the standard deviation of job performance in monetary terms (SDy) is based on the assumption that people are able to estimate the percentiles of a normal distribution. The aim of the research reported here was to test the veracity of this assumption. We used participants who were primed to work with percentiles on a task that provided all the information necessary to solve the problem. Participants' percentile estimates were found to be grossly in error, suggesting that utilities estimated by the Schmidt et al. procedure are inaccurate. This finding was replicated in a second study which also examined the effect of group decision-making on the estimation process. Group estimates were found to be no better than individual estimates.