摘要
通过对2本化学教育核心期刊2011-2024年发表的182篇文献进行文本分析,考察了p值使用中的误用与操控现象及其防范策略。研究发现p值误用主要表现为表述错误、报告不完整、错误选择显著性水平及忽视样本量效应等;p值操控则体现为事后假设设定、未校正的多重比较及选择性报告结果等。针对这些问题,研究提出在研究设计阶段确立方法论规范,建立包含效应量、置信区间的多元统计指标体系,同时完善期刊审查机制。p值应被视为统计推断的组成部分而非唯一标准,研究实践需将统计严谨性与教育实际价值相结合。
Through text analysis of 182 papers published in two core journals of chemical education from 2011 to 2024,the misuse and manipulation of p-values and their prevention strategies were investigated.The study found that the misuse of p-values mainly manifested in incorrect statements,incomplete reports,incorrect selection of significance levels,and neglect of sample size effects;p-hacking manifested in post hoc hypothesis setting,uncorrected multiple comparisons,and selective reporting of results.In response to these problems,the study proposed to establish methodological norms in the research design stage,establish a multivariate statistical indicator system including effect size and confidence interval,and improve the journal review mechanism.The p-value should be regarded as a component of statistical inference rather than the only criterion,and research practice needs to combine statistical rigor with practical educational value.
作者
蓝海航
LAN Hai-Hang(Sun Yat-sen Memorial Middle School,Zhongshan 528454,China)
出处
《化学教育(中英文)》
北大核心
2025年第15期74-79,共6页
Chinese Journal of Chemical Education
基金
2024年度中山市教育科研项目青年课题“具身视角下数字化赋能化学跨学科学习的实践研究”(课题编号:C2024155)
2024年度立项广东省基础教育教研基地(初中化学)项目。
关键词
p值误用
p值操控
效应量
统计推断
研究规范
p-value misuse
p-hacking
effect size
statistical inference
research norms
