In a recent reappraisal of evidence adduced for allocative efficiency in traditional agriculture, Dillon and Anderson employed a method based on a combination of traditional economic aspects and statistical aspects of estimated production functions. However, their reappraisal involved using marginal rather than the (appropriate) joint distributions of the regression coefficients. This paper investigates the error associated with using the inappropriate marginal distribution in a reassessment of allocative efficiency in wool growing in the Australian pastoral zone.