Department of Psychology, University of Connecticut, 406 BabbidgeRd., Unit 1020, Storrs, CT 06269-1020, USA.
J Pers Soc Psychol. 2012 Feb;102(2):408-26. doi: 10.1037/a0025559. Epub 2011 Oct 3.
Though most personality researchers now recognize that ratings of the Big Five are not orthogonal, the field has been divided about whether these trait intercorrelations are substantive (i.e., driven by higher order factors) or artifactual (i.e., driven by correlated measurement error). We used a meta-analytic multitrait-multirater study to estimate trait correlations after common method variance was controlled. Our results indicated that common method variance substantially inflates trait correlations, and, once controlled, correlations among the Big Five became relatively modest. We then evaluated whether two different theories of higher order factors could account for the pattern of Big Five trait correlations. Our results did not support Rushton and colleagues' (Rushton & Irwing, 2008; Rushton et al., 2009) proposed general factor of personality, but Digman's (1997) α and β metatraits (relabeled by DeYoung, Peterson, and Higgins (2002) as Stability and Plasticity, respectively) produced viable fit. However, our models showed considerable overlap between Stability and Emotional Stability and between Plasticity and Extraversion, raising the question of whether these metatraits are redundant with their dominant Big Five traits. This pattern of findings was robust when we included only studies whose observers were intimately acquainted with targets. Our results underscore the importance of using a multirater approach to studying personality and the need to separate the causes and outcomes of higher order metatraits from those of the Big Five. We discussed the implications of these findings for the array of research fields in which personality is studied.
尽管大多数人格研究人员现在都认识到五大特质的评分并非正交的，但该领域对于这些特质相关性是实质性的（即由高阶因素驱动）还是人为的（即由相关测量误差驱动）存在分歧。我们使用元分析多特质多评价者研究来估计控制共同方法偏差后的特质相关性。我们的结果表明，共同方法偏差会大大夸大特质相关性，而一旦得到控制，五大特质之间的相关性就变得相对适中。然后，我们评估了两种不同的高阶因素理论是否可以解释五大特质相关性的模式。我们的结果不支持 Rushton 及其同事（Rushton & Irwing, 2008; Rushton et al., 2009）提出的人格一般因素，而是支持 Digman（1997）的α和β特质因素（由 DeYoung、Peterson 和 Higgins（2002）重新标记为稳定性和可塑性），分别）。然而，我们的模型显示稳定性和情绪稳定性之间以及可塑性和外向性之间存在相当大的重叠，这引发了一个问题，即这些特质因素是否与它们占主导地位的五大特质因素冗余。当我们只包括那些观察者与目标非常熟悉的研究时，这种发现模式是稳健的。我们的结果强调了使用多评价者方法研究人格的重要性，以及需要将高阶特质因素的原因和结果与五大特质的原因和结果分开。我们讨论了这些发现对人格研究领域的一系列研究的影响。
