Predicting adverse impact and mean criterion performance in multistage selection

The authors present an analytical method to assess the average criterion performance of the selected candidates as well as the adverse impact and the cost of general multistage selection decisions. The method extends previous work on the analytical estimation of multistage selection outcomes to the case in which the applicant pool is a mixture of applicant populations that differ in their average performance on the selection predictors. Next, the method was used to conduct 3 studies of important issues practitioners and researchers have with multistage selection processes. Finally, the authors indicate how the method can be integrated into a broader analytical framework to design multistage selection decisions that achieve intended levels of selection cost, workforce quality, and workforce diversity.     

Sequential modulations of stimulus-response correspondence effects depend on awareness of response conflict

In two experiments, sequential modulations of prime—target correspondence effects were investigated in a metacontrast paradigm. Primes were either unmasked and thus consciously discriminable, or entirely masked and thus indiscriminable. Mirroring similar findings from Eriksen- and Simon-type tasks, the influence of prime—target correspondence was reduced in trials that followed a noncorresponding prime—target pair, which suggests that prime-induced response activation can be temporarily suppressed after an incompatible trial. This sequential modulation was independent of prime discriminability in the current trial, but it occurred only when the prime, and thus a conflict between the primeinduced and the deliberately to-be-selected response, was consciously experienced in the preceding trial. This suggests that the suppression of automatic response priming is not an immediate consequence of response conflict, but an intention-mediated strategy.     

A new learning paradigm elicits fast visual discrimination in pigeons

The authors describe a novel multistimulus, multiple-matching learning paradigm for pigeons, which they believe not only simulates pigeons' natural foraging behavior but also more realistically shows their visual discrimination ability. A touch screen thin film transistor panel shows a stimulus configuration consisting of a 2-dimensional array of bitmap images that changes from trial to trial. The pigeon's task during each trial is to peck at images that are defined as positive by the experimenter. An experiment exploring the ability of pigeons to discriminate between achromatic male and female human faces demonstrated the effectiveness of the procedure. In contrast to 12 pigeons that show little improvement over 60 hr of go/no-go discrimination training, 14 multiple-matching pigeons mastered the problem within a few hours.     

Can we shield ourselves from task disturbance by emotion-laden stimulation?

Emotion-laden stimuli can disturb information processing in an unrelated cognitive task. We investigated the possibilities and limitations for shielding from such disturbance. Participants performed a simple categorization task while being simultaneously exposed to negative, neutral, and positive pictures. Performance dropped with negative pictures, relative to positive and neutral stimuli. Unlike Stroop or Simon interference effects, this negativity-based disturbance did not reduce as a function of previous experience of disturbance (Exp. 1) or of announcement of such disturbance on a trial-by-trial basis (Exps. 2 and 3). We found hints of a reduction of negativity-based disturbance, however, when negative stimulation occurred with high list-wide probability (Exp. 4). These observations suggest that the control of negativity-based task disturbance might be possible in a sustained manner, but that it is severely limited when operating in a transient, moment-to-moment manner.     

Stimulus–response bindings contribute to item switch costs in working memory

In counter updating tasks, responses are typically faster when items repeat than when they change (item switch costs). The present study explored the contribution of stimulus–response bindings to these item switch costs. In two experiments, we orthogonally manipulated the repetition/switch of to-be-counted items and the repetition/switch of required manual responses. Item switch costs were considerably lower when item switches were accompanied by response switches than when accompanied by response repetitions. Experiment 2 showed that, although there was also a smaller contribution from stimulus–stimulus bindings (i.e., shape-location), the major part was due to stimulus-response bindings. These results show that in the widely used standard version of the counter updating task, a considerable portion of item switch costs is caused by the unbinding of stimulus–response bindings rather than by processes of switching items in working memory.