MoProc Presentation

The transparency and openness movement, along with research on researcher
analysis practices (i.e., Many Analysts studies) have made it clear that one has many
options for processing and analyzing data. With the increase in computational ability,
researchers can now test multiple versions of an analysis pipeline using multiverse
analysis. Multiverse analysis allows for the investigation of the impact of specific
pipeline choices on the outcome of interest, as well as the potential range of possible
effects given those analysis choices. This workshop will explore multiverse analysis
for research in language sciences, which often deals with the added complexity of
multi-level designs using varied stimuli to assess their phenomenon of interest. The
workshop will provide general examples using priming data and specific R coding
techniques that can be leveraged to complete a multiverse analysis.

透明度与开放性运动，以及关于研究者分析实践（例如，多分析者研究）的研究，已明确指出，在数据处理与分析方面，人们拥有众多选择。随着计算能力的提升，研究者如今能够通过多元宇宙分析来测试分析流程的多个版本。多元宇宙分析使得研究者能够探究特定流程选择对所关注结果的影响，以及基于这些分析选择的潜在效应范围。本次研讨会将探讨语言科学领域中的多元宇宙分析，该领域常需处理使用多样化刺激物评估感兴趣现象的多层次设计所增加的复杂性。研讨会将通过使用前缀数据及特定的 R 语言编码技巧提供一般性示例，这些技巧可用于完成多元宇宙分析。