The role of inhibition of return (IOR) in serial visual search was reinvest
igated using R. Klein's (1988) paradigm of a search task followed by a prob
e-detection task. Probes were presented at either the location of a potenti
ally inhibited search distracter or an empty location. No evidence of IOR w
as obtained when the search objects were removed after the search-task resp
onse. But when the search objects remained on, a pattern of effects similar
to Klein's results emerged. However, when just the search-critical object
parts were removed or when participants received immediate error feedback t
o prevent rechecking of the search objects, IOR effects were observed only
when probes appeared equally likely at search array and empty locations. Th
ese results support the operation of object-based IOR in serial visual sear
ch, with IOR demonstrable only when rechecking is prevented (facilitating t
ask switching) and monitoring for probes is not biased toward search object
s.