Robustness of the adult statistical word segmentation literature: Part 1

We report the first set of results in a multi-year project to replicate every adult statistical word segmentation study. We reported replications of six experiments. The purpose of these replications is both to assess the strength of the findings in the statistical learning literature but also to provide more accurate effect size estimates. In every instance, we were able to replicate successful learning. However, many theoretically important modulations of that learning failed
to replicate. Moreover, learning success was generally much lower than in the original studies. In the General Discussion, we consider whether these differences are due to differences in subject populations, low power in the original studies, or some other factor. Regardless, these initial results suggest taking caution in relying on the originally reported findings.

本报告揭示了历时数年研究项目的初步成果，旨在复现所有成年统计词汇分割研究。我们已对六项实验进行了复现。这些复现的目的不仅在于评估统计学习文献中结论的强度，而且在于提供更精确的效果量估计。在所有情况下，我们都成功实现了学习。然而，许多理论上具有重要意义的对学习过程的调节并未能在复现中得到证实。此外，学习成功率普遍低于原始研究。在总体讨论中，我们探讨了这些差异是否源于受试者群体之间的差异、原始研究中的低效度，或其它因素。无论如何，这些初步结果提示我们在依赖原始报告的发现时需持谨慎态度。