Theories of attention can be separated into those that select by location,
and those that select by location-invariant representation. Experiments dem
onstrating stronger interference or facilitation from distracters grouped b
y nonspatial features with the target than ungrouped distracters have been
considered as evidence for the selection of location-invariant representati
ons. However, few studies have measured spatial attention directly at the l
ocations of the grouped or ungrouped objects. In these experiments subjects
responded to spatial probes (dots) while also identifying a cued target le
tter among distracters. Probe responses were faster for distracter location
s with the target color than for those with the nontarget color, implying t
hat target-color locations receive more attention. This pattern of spatial
attention may explain why target-color distracters interfere more with targ
et identification than nontarget-color distracters. These results suggest t
hat although attention can be directed by nonspatial properties such as gro
uping by color or organization of the scene into objects, selection may ult
imately be based on location. (C) 2001 Elsevier Science Ltd. All rights res
erved.