California Q-sort psychometrics.
California Q-sort psychometrics.
复制标题
加州 Q 排序心理测量学。
DOI：
10.1037/11748-006
复制
发表时间：
2008
影响因子：
2.4
通讯作者：
J. Block
中科院分区：
文献类型：
--
作者：
J. Block
在提供了导致对规定的、对称的、单峰分布的决定的推理之后，下一个问题是"评估者通过既定的CQ程序描述一个人的依赖性如何？"如果在形成CQ排序中涉及的心理操作是重要的任意性，那么整个企业就会失败。首先，需要对评价者的CQ分类的实质性或可复制性作出某种保证。以下是一些偶然的、不可复制的（甚至是有趣的）证据，证明了一个评估者对CQ排序的可复制性。几年前，在一个评估中心，一名研究人员在一系列周末对一组人进行评估，这对被评估者和评估人员来说都是一种折磨。一个周末之后，CQ的每一位工作人员都对刚刚观察到的10个人进行了描述。在一天之内认真地描述10个不同的人是一件令人疲惫的、必然是令人着迷的苦差事。有一次，一名工作人员一时疏忽，完全忘记了他已经对特定主题进行了CQ排序。在漫长的一天快结束时，他第二次对那个人进行了CQ排序，他认为这是他一天中的第10次CQ排序。当他发现自己的重复不必要时，他疲惫地咒骂，因为这意味着他还有最后一个人要描述，然后他认真地进行了CQ分类。这一事件是反常的，不是有意重复的。它是独特的兴趣，因为在纯粹的意义上，它符合理论上，但无法实现的要求，重测信度：可靠性计算没有经过时间的推移，不受残余记忆。当然，这位健忘的工作人员如果还记得他之前几个小时的评估，就不会做第二次（不必要的）CQ排序。此外，这位评估者还经历了许多来自他的各种干预CQ类型的其他被评估者的追溯性抑制。考虑到偶然性，我们可以有把握地假设，健忘的评估者代表了一个评估重测信度的"纯粹"测试案例：他完全不记得在之前的几个小时里，他是如何使用CQ描述符来评估相关人员的。由同一个无记忆的评估者对同一个人进行的两种CQ分类之间的相关性如何？答案是令人印象深刻的高。94.当然，一个如此特殊的例子不应该不谨慎地概括;人们无法知道其他类似健忘的评估者会如何表现。然而，应该记住，记录的主观印象的再现性或可靠性并不一定意味着主观印象具有真实性;个人评估的可靠性并不一定意味着该评估的准确性。个体评价的可验证性取决于更广泛的个体评价者所持有的主观印象的更大的、周围的法则框架。当对同一个人的多个、独立操作的、基于差异的主观印象证明都表现出一定程度的正相关时，个体评价的可解释性就增强了。尽管这绝不是一个万无一失的可信性原则，但来自不同观察者对同一个人的多个独立评价之间的一致性提供了一个有说服力的归纳基础，可以认为他们的重心--一个复合体--具有逼真性，一种真理。长期以来，人们一直认为，从群体组合中得出的判断具有更高的说服力或智慧。经验证明，各种独立观察员之间的一致意见在很大程度上-如果不是总是-是可信的。一个幸福的结果...
Having provided the reasoning that leads to the decision for a prescribed, symmetrical, unimodal distribution, a next question is" How dependably can an appraiser describe a person through the established CQ procedure?" If the mental operations involved in formulating a CQ-sort are importantly aleatoric, the entire enterprise fizzles. First, then, some assurance of the substantiality or replicability of an evaluator's CQ-sort is required. Herewith is some accidental, unreproducible (even amusing) evidence on the reproducibility of a CQ-sort by an appraiser. Some years ago, at an assessment center, a research staff evaluated sets of persons over a series of weekends, grueling for both the assessees and the staff assessors. Immediately after a weekend, each staff member CQ described each of the 10 persons just observed. Seriously characterizing 10 different persons all in the course of a single day is a fatiguing, necessarily absorbing chore. By a slip of the mind, on one occasion a staff member completely forgot that he had already CQ-sorted a particular subject. Toward the end of a long day, he CQ-sorted that person a second time, for what he thought was his 10th CQ-sort of the day. On discovering his unnecessary duplication, he cursed tiredly because it meant he still had a last person to describe, who he then conscientiously CQ-sorted. The incident is anomalous and not planfully repeated. It is of unique interest because in a pure sense it meets the theoretical but unattainable requirement of test-retest reliability: reliability calculated without the passage of time and unaffected by residual memory. Certainly, the forgetful staff member would not have done the second (unnecessary) CQ-sort had he at all remembered his prior, hours-before appraisal. Additionally, this assessor had experienced much retroactive inhibition from his various intervening CQ-sorts of other assessees. It may safely be presumed, given the happenstance, that the forgetful assessor represents a" pure" test case for evaluating test-retest reliability: He had absolutely no memory of how, in the hours just before, he had placed the CQ descriptors in evaluating the person involved. How well did the two CQ-sorts of the same individual by the same but memoryless assessor correlate? The answer is an impressively high. 94. Of course, a single and so special an instance should not be incautiously generalized; one cannot know how other similarly forgetful assessors would have fared. It should be remembered, however, that the reproducibility or reliability of a recorded subjective impression does not necessarily mean the subjective impression has veracity; the reliability of an individual appraisal does not necessarily mean an accuracy of that appraisal. The plausibility of an individual appraisal depends on the larger, surrounding nomological framework of subjective impressions held by a broader base of individual appraisers. The plausibility of an individual appraisal is enhanced when multiple, operationally independent, diversely based subjective impressions of the same person prove to all manifest some sufficient degree of positive intercorrelation. Although by no means a foolproof credibility principle, the congruence among multiple, separate appraisals from diverse observers of the same person offers a persuasive inductive basis for considering their center of gravity--a composite--as having verisimilitude, a kind of truth. Long recognized has been the heightened cogency or wisdom afforded by judgments issuing from group composites. Agreement among a mixed bag of independent observers has proven empirically to be largely--if not always--creditable. A happy consequence of a …