Christine P. Chai’s contribution to the Discussion of ‘Safe testing’ by Grünwald, De Heide, and Koolen
Christine P. Chai 对《安全检验》的讨论AI 解读 ↓
← 返回论文列表与筛选
阅读说明
本页整理公开书目信息，不生成学术质量评分。AI 解读根据题名及所注明的摘要或原文片段生成，未完成全文核验，可能有误；具体假设、证明与结论请核对原文。摘要保留来源文本，未经模型改写。
DOI 首次登记不等于接收或发表日期。部分更正、评论等条目没有独立摘要；缺少可靠材料时保留待补状态。
引用量采用 OpenAlex 单条关联记录，不合计多个记录；关联记录可能包含正式发表版本，不等于此版本单独引用。缺失不视为零，引用指标不代表论文质量。
关注作者按书目中的完整姓名匹配，同名和姓名变体需人工辨别。相关论文先按共享细分方向和具体方法排列，再参考通用方法与发表日期。
分类依据与纠错
AI 根据所提供的论文材料归类：内容属于通用检验方法的讨论与评论，没有直接涉及所列领域细分。
反馈分类问题
正在核对关联版本…
代码、数据与补充材料
正在核对对应版本的资源…
关注这篇论文的作者
Christine P. Chai 对《安全检验》的讨论
- 研究问题
- 这篇讨论关注 e 值与怀疑式 p 值如何比较、可选继续和可选停止如何选择，以及主论文对功效参数的表述是否一致。
- 方法与思路
- 评论者围绕主论文提出比较与应用问题，并建议补充数据收集成本的动机、澄清实验参数和改善图形可读性。
- 材料陈述的贡献
- 所给材料的贡献是提出澄清问题与表达改进建议；其中关于 e 值保证的陈述是在介绍主论文，并非评论者新增的理论结果。
- 阅读关注点
- 应对照主论文和作者回复核对 e 值保证的条件、两种序贯设置的区别及功效记号，区分评论疑问与已经回答的问题。
- 阅读边界
- 这是讨论文本，材料声明没有新增或分析数据。仅据版本未注明的 OpenAlex 索引内容，不能确定主论文实验参数或回复结论。
原文摘要
This paper proposes the e-value for ‘safe testing’ as an alternative to the p-value in hypothesis testing. The e-value allows for optional continuation while preserving Type-I error guarantees. I hope to see lots of discussion at the Royal Statistical Society meeting, like the previous ‘Signs and sizes: Understanding and replicating statistical findings’ (Hutton et al., 2020). I have some questions regarding the descriptions: I see some comparisons between the e-value and several alternatives to the ordinary p-value. How does the e-value compare with the sceptical p-value proposed by Held (2020), which is designed to combine replication studies? This paper discusses both optional continuation (OC) and optional stopping (OS) settings. Are there any real-life examples where OC or OS is more appropriate? If we can choose either one, what are the guidelines to decide? In hypothesis testing, β is the probability of making a type II error, i.e. failing to reject the null when it is actually false (Banerjee et al., 2009). I am confused with the narrative, because some parts say the power is set to 1−β=0.8 and other parts say β=0.8. Which is the one used in the experiments? To further improve the manuscript, I also have some comments: It is common to decide whether to perform additional tests depending on the outcome of previous tests, and the rationale is that we do not want to spend money on more testing unless it is necessary. For better clarification, we can state in the motivation that data collection is expensive (Coffey & Elliott, 2024; Jones et al., 2015). I agree with the consensus that we should be cautious in using p-values for hypothesis testing (Wasserstein & Lazar, 2016). There has also been a recent shift from hypothesis testing to estimation as ‘the new statistics’ (Cumming, 2014; Kruschke & Liddell, 2018), and I hope that e-values is a step to better quantification of the decision-making uncertainty. Finally, it is important to make the images accessible to everyone, including people with disabilities. Please make Figures 2 and 4 more readable by increasing the font size of the axis labels. I also recommend using different line shapes to represent each trend, because people with colour blindness may not be able to see the colours. No new data were generated or analysed in support of this research.