研究方法与开放科学 Psychology

8 个科学问题 · 心理认知 Psychology

本页目录

p 值作为统计显著性指标有哪些局限?

What are the limitations of the p-value as a measure of statistical significance?

p 值只反映数据与零假设的相容程度,并不表示效应大小或实际重要性,也容易受样本量影响。美国统计协会指出,p 值低于阈值不应被当作「真相」的证据,须结合效应量与置信区间解读。
A p-value only gauges compatibility of data with the null, not effect size or practical importance, and is sensitive to sample size. The ASA warns a threshold-crossing p is not evidence of truth and must be read with effect sizes and confidence intervals.
难度 专业 · 适合 高中(15-18) CC BY 4.0 · 来源待核验 · Source pending review
来源 · Source

出处备注:著录仅给书名,无版次与年份;同一书名常对应多个版次,不臆测具体版本。

美国统计协会(ASA) p值声明;《心理学年度评论》方法卷

American Statistical Association (ASA) statement on p-values; Annual Review of Psychology (methods)

#p值 #统计显著性 #效应量 #ASA

纵向研究与横断研究在推断年龄与变化时有何优劣?

What are the trade-offs between longitudinal and cross-sectional designs in inferring age and change?

横断研究在同一时点比较不同年龄群,快速但不易区分年龄效应与世代效应;纵向研究追踪同一群体随时间变化,能分离个体发展轨迹却耗时且面临流失偏倚。二者常被交叉序列设计互补。
Cross-sectional studies compare age groups at one time—fast but confound age with cohort effects; longitudinal studies track the same people, separating developmental trajectories yet costly and prone to attrition bias. Sequential designs combine both.
难度 专业 · 适合 高中(15-18) CC BY 4.0 · 来源待核验 · Source pending review
来源 · Source

出处备注:著录仅给书名,无版次与年份;同一书名常对应多个版次,不臆测具体版本。

《心理学年度评论》(Annual Review of Psychology) 方法卷

Annual Review of Psychology (methods)

#纵向研究 #横断研究 #世代效应 #流失偏倚

预注册如何减少 p-hacking 与选择性报告?

How does preregistration reduce p-hacking and selective reporting?

预注册要求研究者在数据收集前公开研究假设、样本量与分析计划,使事后「数据挖掘」与灵活更改难以伪装成先验假设。它区分了探索性与验证性研究,提升结论可信度,但需配合透明方能见效。
Preregistration requires posting hypotheses, sample size, and analysis plans before data collection, making post-hoc data mining hard to pass as a priori. It separates exploratory from confirmatory work and raises credibility, but only with transparency.
难度 专业 · 适合 高中(15-18) CC BY 4.0 · 来源待核验 · Source pending review
来源 · Source

出处备注:著录仅给书名,无版次与年份;同一书名常对应多个版次,不臆测具体版本。

Nature Human Behaviour 开放科学方法;《心理学年度评论》方法卷

Nature Human Behaviour (open science methods); Annual Review of Psychology (methods)

#预注册 #p-hacking #分析计划 #透明

为什么双盲实验常被视为因果推断的金标准?

Why is the double-blind experiment often the gold standard for causal inference?

双盲设计使受试者与施测者均不知分组,可同时排除受试者的期望效应与研究者的无意偏倚。它在药物与行为干预中最大程度隔离处理变量的净效应,是建立因果的强有力工具。
Double-blind designs withhold group assignment from both participants and experimenters, ruling out expectation effects and unconscious bias. In drug and behavioral trials they isolate the net treatment effect, a strong tool for causality.
难度 进阶 · 适合 初中(12-15) CC BY 4.0 · 来源待核验 · Source pending review
来源 · Source

出处备注:著录含书名,但缺版次 / 出版社 / 年份中的一部分,需编辑补全后才能做独立比对。

沙迪什、库克、坎贝尔 (2002). 《广义因果推断的实验与准实验设计》. Houghton Mifflin

Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and Quasi-Experimental Designs for Generalized Causal Inference. Houghton Mifflin

#双盲 #随机对照 #期望效应 #因果

效应量与置信区间在结果解释中为何重要?

Why are effect sizes and confidence intervals important in interpreting results?

效应量量化关系的强度,置信区间给出估计的精度与不确定性范围,二者弥补了 p 值只回答「是否有差异」的不足。报告效应量有助于跨研究比较与元分析,是结果可解释性的核心。
Effect sizes quantify relationship strength and confidence intervals convey precision and uncertainty, complementing p-values that only ask whether a difference exists. Reporting them aids cross-study comparison and meta-analysis, central to interpretable results.
难度 进阶 · 适合 初中(12-15) CC BY 4.0 · 来源待核验 · Source pending review
来源 · Source

出处备注:著录仅给书名,无版次与年份;同一书名常对应多个版次,不臆测具体版本。

《心理学年度评论》(Annual Review of Psychology) 方法卷;美国心理学会(APA)

Annual Review of Psychology (methods); American Psychological Association (APA)

#效应量 #置信区间 #统计解释 #元分析

心理学领域的「可重复危机」指的是什么?

What is the replication crisis in psychology?

2010 年代多项大型重复实验发现,许多经典与高影响力研究无法被成功复制,效应显著缩水。这暴露了 p-hacking、灵活分析与发表偏倚等问题,推动了开放科学改革。
In the 2010s large replication projects found many classic and high-impact findings failed to reproduce, with sharply reduced effects. This exposed p-hacking, flexible analysis, and publication bias, spurring open-science reform.
难度 进阶 · 适合 大众 CC BY 4.0 · 来源待核验 · Source pending review
来源 · Source

出处备注:著录仅给书名,无版次与年份;同一书名常对应多个版次,不臆测具体版本。

《心理学年度评论》(Annual Review of Psychology) 方法卷;Nature Human Behaviour

Annual Review of Psychology (methods); Nature Human Behaviour

#可重复危机 #p-hacking #发表偏倚 #开放科学

开放科学中的数据共享实践有哪些价值与顾虑?

What are the benefits and concerns of data-sharing practices in open science?

数据共享提升研究透明度、可重复性与二次分析效率,并减少选择性报告。其顾虑包括参与者隐私保护、数据误用与弱势群体风险,因此需在伦理审查与去标识化框架下有序推进。
Data sharing boosts transparency, reproducibility, and reuse while reducing selective reporting. Concerns include participant privacy, misuse, and risks to vulnerable groups, so it must proceed under ethics review and de-identification.
难度 进阶 · 适合 初中(12-15) CC BY 4.0 · 来源待核验 · Source pending review
来源 · Source

出处备注:著录仅给书名,无版次与年份;同一书名常对应多个版次,不臆测具体版本。

Nature Human Behaviour 开放科学政策;《心理学年度评论》方法卷

Nature Human Behaviour (open science policy); Annual Review of Psychology (methods)

#数据共享 #开放科学 #可重复性 #隐私

元分析在心理学研究中起什么作用?

What role does meta-analysis play in psychological research?

元分析通过系统整合多项独立研究、用统计方法估计整体效应,能发现单个小样本研究难以检出的稳定模式,并量化异质性来源。它是证据分级的顶层方法,但依赖纳入研究的质量。
Meta-analysis systematically combines independent studies to estimate an overall effect, revealing stable patterns small samples miss and quantifying heterogeneity. It tops evidence hierarchies but depends on the quality of included studies.
难度 进阶 · 适合 初中(12-15) CC BY 4.0 · 来源待核验 · Source pending review
来源 · Source

出处备注:著录仅给书名,无版次与年份;同一书名常对应多个版次,不臆测具体版本。

《心理学年度评论》(Annual Review of Psychology) 方法卷

Annual Review of Psychology (methods)

#元分析 #证据整合 #异质性 #效应量