Abstract

While research has documented the framing effects of performance data on decision-making, little is known about the impact of data visualizations on performance gaming. We propose that performance information portraying prosocial impact reduces gaming if the information is episodic and visualizes benefits for clients in need. We also theorize about expert-novice differences and suggest that performance information is less influential among experts. We conduct an experiment with samples of citizens (novices) and school leaders (experts) in which subjects must decide whether they are willing to omit poorly performing students when calculating school-wide performance scores. We find that school leaders game less than citizens, and that the former are less responsive to episodic information cues than the latter. However, information becomes more influential if it can capture school leaders’ attention and if leaders have little work experience. More broadly, our findings suggest being cautious with generalizations across groups of data users.

摘要

虽然研究记录了绩效数据对决策的框架效应，但对数据可视化对绩效游戏的影响知之甚少。我们建议，如果信息是情节性的并且为有需要的客户提供了可视化的好处，那么描绘亲社会影响的绩效信息会减少游戏。我们还对专家与新手的差异进行了理论分析，并建议性能信息对专家的影响较小。我们以公民（新手）和学校领导（专家）为样本进行了一项实验，在该实验中，受试者必须决定他们是否愿意在计算全校表现分数时忽略表现不佳的学生。我们发现学校领导的游戏比公民少，并且前者对情节信息线索的反应不如后者。然而，如果信息能够吸引学校领导的注意力并且领导者没有多少工作经验，它就会变得更有影响力。更广泛地说，我们的研究结果表明，对跨数据用户群体的概括要谨慎。