Perceptual interference is the finding that prior exposures to a target item's partial features inhibit its later identification. This study tests a competitive activation model of perceptual interference that attributes interference to the activation of competing responses generated by prior cues. We examined 2 sets of data that seemed to be inconsistent or incompatible with the model. The first is the observed positive effect of viewing time of stimuli (J. S. Bruner & M. C. Potter, 1964). The second is the finding that interference occurs only for studied or primed words (Z. F. Peynircioglu, 1987; Z. F. Peynircioglu & M. J. Watkins, 1986). Experiments 1 and 2 showed paradoxical effects of viewing time and found evidence supporting the competitive activation model. Experiments 3–6 failed to replicate Peynircioglu and Watkins' finding and showed that a performance level explanation compatible with the competitive activation model can account for all related results.