McDowell C, Acklin M W
Honolulu, Hawaii, USA.
J Pers Assess. 1996 Apr;66(2):308-20. doi: 10.1207/s15327752jpa6602_9.
Although the Rorschach test has demonstrated significant refinements in reliability, validity, and statistical power as a result of the procedural standardization and scoring innovations introduced by Exner's Comprehensive System, the issue of Rorschach interrater reliability remains unexplored. This article examines the psychometric foundations of Rorschach interrater reliability and applies notions from applied behavioral analysis to the treatment of Rorschach data. We empirically compare 3 methods of quantifying interrater agreement, their accuracy in estimating interrater agreement, and efficiency in reducing error in Rorschach research. Results indicate that the magnitude of differences between methods of quantifying interrater agreement and the associated reductions of error are significant. We propose a standard method for quantifying interrater agreement in Rorschach research.
尽管由于埃克纳综合系统引入的程序标准化和计分创新，罗夏墨迹测验在信度、效度和统计效力方面有了显著改进，但罗夏测验评分者间信度的问题仍未得到探讨。本文考察了罗夏测验评分者间信度的心理测量学基础，并将应用行为分析的概念应用于罗夏测验数据的处理。我们实证比较了3种量化评分者间一致性的方法、它们在估计评分者间一致性方面的准确性以及在减少罗夏测验研究误差方面的效率。结果表明，量化评分者间一致性的方法之间的差异程度以及相关的误差减少是显著的。我们提出了一种在罗夏测验研究中量化评分者间一致性的标准方法。
