术语“评分者样本”的详细信息

英译术语:

scorer sample

汉语语境:

概化理论是一种重要的信度理论，依据测验维度的多少，可分为一元概化理论和多元概化理论。在人才招聘中，常用无领导小组讨论和行为面试两种技术对候选人的综合能力进行测评。为研究测试中的评分者样本信度，研究应用一元交叉设计和多元交叉设计对某公司的某次人事测评结果的评分者样本的一致性信度进行评估。研究结果表明，两种模式下的信度系数都达到了较高水平，相比较而言，一元交叉设计信度低于多元交叉设计的合成概化系数。在评分者个数增加到4-6个时，测试的评分者信度会有较大的提高。

英语语境:

Generalizability theory is an important reliability theory. Based on the number of test dimensions, generalization theory can be divided into univariate generalizability theory and multivariate generalizability theory. In employee recruitment, leaderless group discussion and structured interview are two techniques often applied to the evaluation of candidates'integrated capabilities. To study the scorer sample'reliability, this research adopted two measurement designs, univariate and multivariate single facet crossed designs. The results showed that these two designs both had high reliability coefficients; comparatively speaking, the reliability coefficients of the univariate crossed design came below the integrated generalizability coefficients of the multivariate crossed design. When the number the raters increased to 4-6, the scorer sample'reliability would greatly improved.