结果已知后假设 [HARKing]
定义: 结果已知后假设（Hypothesizing After the Results are Known，简称HARKing），是一种存在争议的科研实践，其定义为“在研究报告中，把基于研究结果的事后假设，即受结果启发或调整后的假设，伪装成预先提出的先验假设”（Kerr, 1998, p. 196）。例如，研究者可能先进行亚组分析，发现某一亚组存在显著效应后，在论文引言部分提出与该结果相匹配的“假设”。
相关术语: Analytic Flexibility, Confirmatory analyses, Exploratory data analysis, Fudging, Garden of forking paths, P-hacking, Questionable Research Practices or Questionable Reporting Practices (QRPs)
参考文献:
- Kerr, N. L. (1998). HARKing: Hypothesizing after the results are known. Personality and Social Psychology Review, 2(3), 196–217. https://doi.org/10.1207/s15327957pspr0203_4
- Nosek, B. A., & Lakens, D. (2014). Registered reports. Social Psychology, 45, 137–141. https://doi.org/10.1027/1864-9335/a000192
原稿作者: Beatrix Arendt
审阅者: Matt Jaquiery, Charlotte R. Pennington, Martin Vasilev, Flávio Azevedo
翻译者: AI-driven translation tool "TransFlow" (developed by Jinbiao Yang and COSN OpenTransfer team)
译稿审阅者: Xuejun Ryan Ji, Liangjie Chen, Ruoting Liu, Shuxian Jin