Forty-four preschoolers completed 2 conditions of a Stroop-like procedure (e.g., saying "boat" for car and "car" for boat) that differed in whether a 3-s delay was imposed before responding. The test card was visible during the delay period for half of the children and occluded for the other children. Preschoolers' interference control was significantly improved in the delay condition. There was no difference between the two delay variants (test card visible or occluded). Children were more prone to interference as testing progressed regardless of whether the delay was present. These results suggest that delays effectively reduce interference by reducing the potency of the competing response during test trials, although memory demands may moderate the effectiveness of delays. (Contains 3 tables.)