Rogers R, Sewell K W, Morey L C, Ustad K L
Department of Psychology, University of North Texas, Denton 76203-6587, USA.
J Pers Assess. 1996 Dec;67(3):629-40. doi: 10.1207/s15327752jpa6703_15.
Psychological assessment with multiscale inventories is largely dependent on the honesty and forthrightness of those persons evaluated. We investigated the effectiveness of the Personality Assessment Inventory (PAI) in detecting participants feigning three specific disorders: schizophrenia, major depression, and generalized anxiety disorder. With a simulation design, we tested the PAI validity scales on 166 naive (undergraduates with minimal preparation) and 80 sophisticated (doctoral psychology students with 1 week preparation) participants. We compared their results to persons with the designated disorders: schizophrenia (n = 45), major depression (n = 136), and generalized anxiety disorder (n = 40). Although moderately effective with naive simulators, the validity scales evidenced only modest positive predictive power with their sophisticated counterparts. Therefore, we performed a two-stage discriminant analysis that yielded a moderately high hit rate (> 80%) that was maintained in the cross-validation sample, irrespective of the feigned disorder or the sophistication of the simulators.
使用多尺度量表进行心理评估在很大程度上依赖于被评估者的诚实和坦率。我们调查了人格评估量表（PAI）在检测假装患有三种特定障碍（精神分裂症、重度抑郁症和广泛性焦虑症）的参与者方面的有效性。通过模拟设计，我们在166名天真的参与者（准备很少的本科生）和80名经验丰富的参与者（有1周准备时间的心理学博士生）身上测试了PAI效度量表。我们将他们的结果与患有指定障碍的人进行比较：精神分裂症患者（n = 45）、重度抑郁症患者（n = 136）和广泛性焦虑症患者（n = 40）。虽然效度量表对天真的模拟者有一定效果，但对经验丰富的模拟者而言，其阳性预测力仅为中等水平。因此，我们进行了两阶段判别分析，得到了较高的命中率（> 80%），且在交叉验证样本中保持不变，无论假装的障碍类型或模拟者的经验如何。
