Memory Suppression Does Not Slow Down, but Retrieval Practice Speeds Up Later Recall in the Think/No-Think Task

In the Think/No-Think (TNT) task, the effect of memory suppression on later recall is typically assessed by the difference in total recall score between No-Think and baseline words, termed suppression-induced forgetting (SIF). Considering the usually small, and occasionally null, effects, we reason that this recall score, that is based on a binary measure (remembered or forgotten), might be too insensitive to detect the effects of suppression on later recall. In this study, we therefore tested recall delay as a potentially more sensitive measure, as it is measured on a continuum. Recent studies using modified TNT tasks suggest that No-Think words are indeed recalled slower than baseline or Think words. We tested this effect using a typical TNT task with word pairs, in a larger sample (N = 80), and with a preregistered analysis plan complemented by Bayesian analyses. Moreover, we tested recall delay for emotionally negative words. Using a same-probe recall test, the SIF effect was significant but small, and it was not supported by a Bayesian analysis. Moreover, the SIF effect was not significant for neutral and negative words separately. Retrieval suppression did not slow down recall time, but retrieval practice (i.e., Think condition) sped up later recall compared to baseline words. These findings support growing concerns regarding the robustness of suppression effects on later recall.