Perceptual load theory (Lavie, 1995) claims that visual selection is determined both by the perceptual demands a display imposes and by the perceptual resources an observer has available for processing. This theory is often tested by examining distractor interference in modified flanker tasks, which separate potential target locations from distractor locations and allow researchers to measure distractor processing. Although this task has provided significant insight into cognitive processing, it may also be an example of how a given task obscures potentially important experimental factors. Specifically, the structure of the target array could encourage observers to adopt a narrow attentional window, which could eliminate distractor interference as significant distractor processing has been shown to occur inside but not outside the attentional window (Belopolsky, Zwaan, Theeuwes, & Kramer, 2007). The present experiments included conditions that allowed the target to vary among the same locations as within a structured target array but varied possible nontarget locations that never overlapped with possible target locations. Whenever nontarget items remained in fixed locations, significant distractor processing occurred due to the observer adopting a wide attentional window even under high perceptual load conditions (Experiment 1). Further evidence showed that familiarity with these locations could not explain the interference (Experiment 2A). However, imposing a ring around the target array, similar to what a circular target array might impose, narrowed the attentional window and eliminated distractor interference (Experiment 2B). Thus, the size of the attentional window is capable of dominating both perceptual load and familiarity in visual selection.