Abstract: It is well established in the visual search literature that the time to detect a target increases as the display size increases (provided that target and distractor share some feature). Yantis and Jonides (1984) compared the display size effect in two conditions: onset condition, in which the target (a letter) appeared as an abrupt onset (the distractor letters appeared as changes in previous figure-eight place holders), and non-onset condition, in which the target also appeared as a change in one of the place holders. The size effect in the onset condition was reduced to the usual parallel search parameters. Yantis and Jonides argued that when the target is an onset, it is treated as a new object which captures attention. This item is then the first to be analysed, and so the response time is relatively independent of the number of distractors. Gibson (1996), however, argues that the effect is due to masking of the distractors in the onset condition. In two series of experiments we investigate this issue. When masking problems are eliminated, if the new-object appears as the old-objects reappear, we find the opposite result to Yantis and Jonides: a disadvantage for new-objects. We explain this effect by assuming that in the new-object condition a new object-representation must be created, whereas the old-object representations have only to be updated. However, when the new-object appears as an abrupt onset, responses in the new-object condition are faster than in the old-object condition. Further research is necessary to investigate which properties of new-objects do capture attention.

(1997). UWB Psychology Mini-Conference