Observers require less time to identify a visual target when its location is cued in advance than when it is not cued, and the magnitude of the improvement depends on the validity of the cue. According to J. Jonides's (1983) 2-process model, there exist 2 possible modes of attentional readiness: a focused-attention mode and a diffuse-attention mode. Observers are assumed to enter the focused-attention mode on a proportion of trials that matches the validity of the cue and to enter the diffuse-attention mode on the remaining trials. The present experiment tested and rejected the response time mixture prediction of the 2-process model. An instance of the class of 1-process models in which perceptual objects are sampled in parallel according to the validity of the cue was evaluated. A stochastic simulation of the model yielded results that paralleled those of the experiment.