The effects of aggregation across and within occasions on consistency, specificity and reliability.
The effects of aggregation across and within occasions on consistency, specificity and reliability.
复制标题
跨场合和场合内的聚合对一致性、特异性和可靠性的影响。
DOI：
--
复制
发表时间：
1990
期刊：
影响因子：
--
通讯作者：
Manfred J. Schmitt
中科院分区：
文献类型：
--
作者：
R. Steyer;Manfred J. Schmitt
摘要：经常提倡聚合来提高行为测量的一致性并降低行为测量的情境特异性。据说，这种程序可以通过经典心理测量测试理论来证明，特别是用于延长测试的可靠性的斯皮尔曼-布朗公式。然而，经典心理测量测试理论忽略了情境的系统影响。因此，它不能作为讨论情境特异性的基础。作为替代方案。我们提出对经典心理测量测试理论的概括，以使得区分可靠性、一致性和特异性系数。在对不同情况下的人员进行并行测量的模型的基础上，研究了一个特殊的因素模型，并且可以从总体和非总体变量的相关性中确定可靠性、一致性和特异性系数。变量之间的一致性并不一定代表高度一致性。
"' Abstract: Aggregation has often been advocated to increase consistency and to decrease situational specificity of behavioral measures. Perportedly, such a procedure can be justified by classical psychometric test theory, especially the Spearman-Brown formula for reliability oflenghtened tests. Classical psychometric test theory. however. ignores systematic effects of situations. Therefore, it cannot serve as a basis for discussions of situational specificity. As an alternative. we propese a generalization of classical psychometric test theory to make possible the distinction between reliability, consistency. and specificity coefficients. On the basis of a model of parallel measurements of persons-in-situations. the implications of aggregation within and across occasions for consistency, specificity, and reliability are investigated. The model implies a special factor model and the coefficients of reliability, consistency, and specificity may be identified from the correlations of the aggregate and nonaggregate variables. Two of the most surprising results of our investigation are: (a) there are conditions in which specificity may increase as a consequence of aggregation; (b) a high correlation between variables is not necessarily indicative of high consistency.
聚合是否会产生虚假的高行为稳定性估计？
影响因子：
7.6
作者：
Epstein,S
通讯作者：
Epstein,S