State-Trace Analysis Meets Personality Measurement: Identifying and Fixing Hidden Inconsistencies in the Big Five Questionnaires

Factor analysis falls short in addressing a pivotal question within personality measurement: the determination of whether a set of items can be logically reduced to a single latent factor. This study advocates for the application of state-trace analysis, an underutilized method from mathematical psychology, to decisively address this question. State-trace analysis introduces a simple but rigorous criterion for unidimensionality: monotonicity between item pairs. Identification of items violating this criterion within a factor is straightforward. This paper illustrates exemplary analyses within the framework of the five-factor model, focusing on the International Personality Item Pool-NEO-120 ($N=618,000$) and the NEO Personality Inventory--Revised (N₁=857, N₂=500) questionnaires. The findings demonstrate that sustaining a five-factor model necessitates alterations to many items. This underscores the potency of state-trace analysis in advancing personality measurement beyond current methodologies. The paper concludes by discussing strategies to promote broader adoption of this method and how future designs in personality research can be tailored to effectively incorporate state-trace analysis. notReviewed other

因素分析（factor analysis）在处理人格测量领域的一项核心问题时存在局限：即判定一组量表条目能否在逻辑上被简化为单一潜在因子(latent factor)。本研究主张应用来自数学心理学领域的一种尚未得到广泛使用的方法——状态轨迹分析(state-trace analysis)，以彻底解决这一问题。状态轨迹分析为单维性(unidimensionality)提出了一项简洁却严谨的判定准则：条目对之间的单调性(monotonicity)。识别某一因子中违反该准则的条目，过程简便直接。本文在五因子模型(five-factor model)的框架下展示了示范性分析案例，重点聚焦于国际人格条目池-NEO-120（样本量$N=618,000$）以及修订版NEO人格问卷(NEO Personality Inventory--Revised，样本量分别为$N_1=857$、$N_2=500$)。研究结果显示，要维持五因子模型的合理性，需对大量条目进行修订。这凸显了状态轨迹分析在推动人格测量方法突破现有局限方面的卓越效能。本文最后讨论了推动该方法获得更广泛应用的策略，以及人格研究领域的未来设计应如何调整，以有效融入状态轨迹分析。未审阅其他内容。