Behaviour occurs not as isolated incidents, but within an ongoing sequence of events. The task-switching paradigm provides a useful way to investigate the impact of different events upon subsequent performance. An implication of two-stage task-switching models is that preparing a task without performing it might affect task readiness only to a limited extent. However, recent research has surprisingly shown larger switch costs following preparation (“cue-only” trials) than following performance (“completed” trials). We set out to conduct a rigorous comparison of the size of switch costs following cue-only versus completed trials. In Experiments 1 and 2, we controlled the timing between critical trial events. This had the effect of roughly equating, but not reversing, the relative size of switch costs. In Experiment 3, we restructured the paradigm to equate the predictability of cue and target events. Switch costs following cue-only trials were now smaller than those following completed trials. These studies confirm that task preparation alone is sufficient to drive subsequent switch costs. They also indicate that task performance might increase the size of these costs, consistent with two-stage task-switching models. Switch costs appear to be affected by both the timing and predictability of trial events.
Bibliographical noteOpen Access via Springer Compact Agreement.
The authors would like to thank Suzanne Wilson for help with data collection (Expt. 1). Data were analysed and graphed using R (R Core Team, 2013). R packages used were as follows: afex (Singmann, Bolker, Westfall, and Aust, 2017); BayesFactor (Morey and Rouder, 2018); dplyr (Wickham and Francois, 2016); ggplot2 (Wickham, 2009); reshape (Wickham, 2007); tidyr (Wickham, 2017).
Experiments 2 and 3 were funded by a small grant from the Experimental Psychology Society, which provided bursaries for KK and AR as well as participant payments.
- COGNITIVE CONTROL
- RESPONSE SELECTION