Department of Mental Health Law & Policy, Florida Mental Health Institute, University of South Florida.
Personal Disord. 2013 Jan;4(1):1-14. doi: 10.1037/a0026482. Epub 2012 Jan 23.
A concern among researchers is that self-report measures may not be valid indicators of psychopathic traits due to the core features of psychopathy (e.g., lying, deception/manipulation). The current study addresses this issue by combining effects sizes from studies published on or before March 31, 2010 to examine the relation between scores of 3 widely used self-report psychopathy measures--the Psychopathic Personality Inventory (PPI; Lilienfeld & Andrews, 1996) and its revised version (PPI-R; Lilienfeld & Widows, 2005) and Levenson's Self-Report Psychopathy scale (LSRP; Levenson, Kiehl, & Fitzpatrick, 1995) and scores on measures assessing response style (i.e., faking good and faking bad). Effect sizes were obtained from 45 studies for total, Factor 1, and Factor 2 scores (faking good: k = 54, 55, and 55, respectively; faking bad: k = 51, 50, and 50, respectively). Based on a random effects model, a significant negative association was found between social desirability/faking good and both total (r(w) = -.11, p < .01) and F2 (r(w) = -.16, p < .01) scores, and moderation analyses suggested that effect sizes varied as a function of psychopathy scale and validity scale used. Significant positive associations were also found between faking bad and both total (r(w) = .27, p < .05) and F2 (r(w) = .32, p < .05) scores. Also, moderation analyses suggested that effect sizes varied as a function of study location, psychopathy scale, and validity scale. Despite several limitations (e.g., inclusion of only published studies, limited moderators, exclusion of other measures), the general findings temper concerns of positive response bias and underscore the validity of self-report psychopathy scales.
研究人员关注的一个问题是，由于精神病态的核心特征（例如，说谎、欺骗/操纵），自我报告的测量方法可能不是精神病态特征的有效指标。本研究通过结合截至 2010 年 3 月 31 日之前发表的研究的效应大小来解决这个问题，以检验 3 种广泛使用的自我报告精神病态量表的得分之间的关系——精神病态人格量表（PPI；Lilienfeld 和 Andrews，1996 年）及其修订版（PPI-R；Lilienfeld 和 Widows，2005 年）和 Levenson 的自我报告精神病态量表（LSRP；Levenson、Kiehl 和 Fitzpatrick，1995 年）以及评估反应风格的量表（即，装好人和好装坏）的得分。从 45 项研究中获得了总得分、因子 1 和因子 2 得分的效应大小（装好人：k=54、55 和 55；装坏人：k=51、50 和 50）。基于随机效应模型，发现社交期望/装好人与总分（r(w)=-.11，p<.01）和 F2（r(w)=-.16，p<.01）得分之间存在显著的负相关，并且调节分析表明，效应大小随精神病态量表和效度量表的使用而变化。还发现装坏人与总分（r(w)=.27，p<.05）和 F2（r(w)=.32，p<.05）得分之间存在显著正相关。此外，调节分析表明，效应大小随研究地点、精神病态量表和效度量表的变化而变化。尽管存在一些局限性（例如，仅包括已发表的研究、有限的调节变量、排除其他量表），但总体发现减轻了对积极反应偏差的担忧，并强调了自我报告精神病态量表的有效性。
