It has become orthodox in economics research to interpret the association between hourly earnings and working hours as the expression of the preferences of workers. This convention originated in H. Gregg Lewis' explanation for the decline in hours of work since the nineteenth century. His explanation rested on an explicit resolution of the identification problem inherent in any quantity (hours) - price (wage) relation. For over forty years, researchers have neglected this identification problem with the result that the findings in the purported "labor supply" literature are of questionable value.