Schiltenwolf, Moritz
Kiesel, Andrea
Frings, Christian
Dignath, David
Funding for this research was provided by:
Deutsche Forschungsgemeinschaft (DI2126/3-1)
Eberhard Karls Universität Tübingen
Article History
Received: 4 October 2022
Accepted: 10 August 2023
First Online: 24 August 2023
Declarations
:
: The authors declare that they have no conflict of interest.
: The present study fulfils the requirements of the generalized ethical approval by the <i>Ethics Committee for Psychological Research</i>, University of Tübingen, Germany.
: The registrations for all experiments can be found on OSF (ExternalRef removed). We deviated from the preregistered analysis plans for two reasons. First, for Experiments 1 and 2, we preregistered a frequentist approach (repeated-measures ANOVA) but decided to switch to a Bayesian framework. This allowed us to compare evidence for a model including an effect of temporal decay on the c-CSE against a null model, and we avoided violations of <i>null hypothesis significance testing</i> assumptions by repeated testing in the mega-analysis. For experiments 3 and 4, we preregistered to test our main hypothesis with a Bayesian dependent measures <i>t</i>-test contrasting difference scores (c-CSE in long ITIs vs. c-CSE in short ITIs). Upon a reviewer’s suggestion, we decided to switch from a Bayesian <i>t</i>-test to a Bayesian ANOVA that includes by-participant random intercepts since this more closely followed the preregistered analysis plans for experiments 1 and 2 and allowed us to test for other effects of interest, such as CEs, CSEs, and c-CSEs, in a single analysis. This analysis protocol was preregistered for experiment 5 and applied for all other individual analyses as well as for the mega-analysis. The applied model adhered closely to the originally preregistered analysis plan using the same factors and dependent variables as well as including by-participant random effects. Bayes factors resulting from the Bayesian ANOVA turned out to be more conservative than Bayes factors from the Bayesian <i>t</i>-tests, which were used for the stopping rule in data collection (see footnote on p. 9). The results of the preregistered analysis plan can be found in the online supplement.