Partial repetition costs persist in nonsearch compound tasks: evidence for multiple-weighting-systems hypothesis

Search performance is sequence-dependent. A specific finding observed in compound-search tasks consists of an interaction between cross-trial sequences (repetition vs. change) of the target-defining (primary) and response-defining (secondary) features: The effect of a target change is greater when the response stays the same than when the response changes. The present study tested whether this interaction arises from processes involved in target search or from later processes in compound tasks. Uncertainty about the upcoming target location-that is, the search component of compound tasks-was removed in different experiments, either by the use of exogenous spatial precues or by presenting only one, central item. Despite having removed the search component, we observed a robust interaction between target (primary) and response (secondary) feature sequences. These results suggest that this interaction originates from a processing stage concerned with discriminating the response feature of a single (selected) item, rather than from a search-related stage. Furthermore, the results support our multiple-weighting-systems hypothesis, according to which sequence effects in visual search tasks do not stem from a single, unitary mechanism; rather, multiple stages of processing on any given trial can lead to separate memory traces, which in turn have effects on different stages of processing on the subsequent trial.