A peripheral red distractor hampers central target identification when the target is defined by red (Folk, Leber, & Egeth, 2002). On the other hand, a peripheral onset distractor fails to affect task performance when the target location is pre-cued by a 100%-valid central cue (Theeuwes, 1991; Yantis & Jonides, 1990). Different paradigms were used in these two sets of studies (RSVP in the former and spatial cueing in the later studies) and this may affect the deployment of attention and the results differently. We asked the participants to search for a red target as in Folk et al. (2002) and compared attentional capture by red distractor (i.e., contingent on the target-defining feature) and onset distractor (i.e., not contingent on the target-defining feature), using the RSVP paradigm (Experiment 1) and the spatial cueing paradigm (Experiment 2 and 3). Results showed that, regardless of which paradigm is used, attentional capture occurs only by the red distractor but not by the onset distractor, suggesting that the different results obtained in previous studies were not caused by different paradigms used. Outside focal attention a salient stimulus that is contingent on the target-defining feature has the highest possibility to capture attention.