New Techniques for Evaluating Temporal Consistency
提出一组新统计量，用于评估测试的时间一致性（重测信度），弥补传统总分相关法忽略项目、项目间关系和个体信息的不足，并通过示例验证其有效性。
Temporal consistency (otherwise known as test-retest reliability) is often calculated by correlating test scores for participants who have taken a test on two separate occasions. Unfortunately, analysis of the relationship between total scores across test administrations potentially ignores information related to the consistency of individual items, item interrelationships, and individual respondents. The purpose of this article is to introduce a set of new statistics intended to address these issues. Illustrative examples are provided to demonstrate the potential utility of each proposed statistic and provide proof of concept. Each of the proposed statistics works as intended, providing evidence that this set of analyses can be used to enhance the assessment of temporal consistency.