The usefulness of half-list speech discrimination tests for evaluating hearing aids was assessed. Test and retest speech discrimination scores (SDSs) were obtained from 23 adults with moderate hearing impairment using four preselected hearing aids. Reliability coefficients were generally lower for half-lists than for full-lists, and the rankings of the hearing aids on the basis of SDSs were less consistent for the half-lists than for the full-lists. Individual subject differences between half-list SDSs were 10% or greater in more than one third of the cases. It was concluded that half-lists derived from the NU-6 full-lists should not be used to differentiate among hearing aids.