When searching for a target in a scene, previously learned spatial context can facilitate response times (Chun & Jiang, 1998). However, it has been debated whether this is driven by attentional guidance (Chun, 2000; Geyer et al., 2010) or response selection processes (Kunar et al., 2007; Kunar & Wolfe, 2011). To address this issue, 18 subjects performed a visual search task in which (a) the spatial configuration was either repeated or novel, and (b) the target was either present or absent. The present/absent manipulation permitted signal detection computations to decouple perceptual sensitivity (d’) and response bias (β). Average RT was faster (p < 0.01) and d’ was increased (p < 0.03) for repeated displays, but there was no effect on β (p > 0.45). These results suggest that early perceptual processes, as opposed to response processes, may provide a greater contribution to the benefits of repeated context in visual search.