Effects of visualizing statistical information - an empirical study on tree diagrams and 2 × 2 tables.
Effects of visualizing statistical information - an empirical study on tree diagrams and 2 × 2 tables.
复制标题
DOI：
10.3389/fpsyg.2015.01186
复制
发表时间：
2015
影响因子：
3.8
通讯作者：
Bruckmaier G
中科院分区：
文献类型：
--
作者：
Binder K;Krauss S;Bruckmaier G
在他们的研究文章中，学者们经常使用包含自然频率的2 × 2表格或树图，以便向同行说明贝叶斯推理情况。有趣的是，到目前为止，这些可视化对参与者表现的影响还没有经过经验检验（除了外显训练研究）。在本文中，我们报告了一项实证研究（3 × 2 × 2设计），在该研究中，我们系统地改变了两种背景（医学背景与经济背景;不是感兴趣的因素）的可视化（无可视化与2 × 2表格与树形图）和信息格式（概率与自然频率）。N = 259名参与者（16-18岁的学生）中的每一个都必须解决两个典型的贝叶斯推理任务（“乳房X线摄影问题”和“经济问题”）。假设2 × 2表格和树形图--特别是当自然频率包括在内时--可以促进对贝叶斯推理情况的众所周知的困难结构的洞察。与许多其他可视化（例如，图标数组、欧拉图）、2 × 2表格和树图具有构造简单的优点。我们的研究结果教学贝叶斯推理的影响将进行讨论。
In their research articles, scholars often use 2 × 2 tables or tree diagrams including natural frequencies in order to illustrate Bayesian reasoning situations to their peers. Interestingly, the effect of these visualizations on participants’ performance has not been tested empirically so far (apart from explicit training studies). In the present article, we report on an empirical study (3 × 2 × 2 design) in which we systematically vary visualization (no visualization vs. 2 × 2 table vs. tree diagram) and information format (probabilities vs. natural frequencies) for two contexts (medical vs. economical context; not a factor of interest). Each of N = 259 participants (students of age 16–18) had to solve two typical Bayesian reasoning tasks (“mammography problem” and “economics problem”). The hypothesis is that 2 × 2 tables and tree diagrams – especially when natural frequencies are included – can foster insight into the notoriously difficult structure of Bayesian reasoning situations. In contrast to many other visualizations (e.g., icon arrays, Euler diagrams), 2 × 2 tables and tree diagrams have the advantage that they can be constructed easily. The implications of our findings for teaching Bayesian reasoning will be discussed.