摘要
信度是测验编制中不可缺的重要指标。使用已有测验开展心理问题研究或临床应用是心理研究中常用的方法和手段。测验手册的信度系数能否引入当前的研究?显然是一个有意义的问题。信度概化是以概化理论的思想及方法指导信度研究的元分析技术,这种分析技术的基本思想是:在承认测验手册信度系数基础上,重点分析测验使用时的信度系数及其变异性,以统计分析方法找出变异性产生的预测源,进而说明测验更适合的样本群体的属性和特征。它有助于研究者更为准确、有效地使用测验,以获得可靠、有效的测验分数,提升测验使用的科学性,避免"信度引入"的错误。
Reliability is the most important index in making a test plan.Using ready-made tests is a common method in researching psychological issues or clinical applications.Can the reliability co-efficiently provided by test 's manual introduce the current research? Apparently,it is a meaningful matter.Reliability generalization(RG) is a meta-analytical technique that is guided by the idea and method of generalization theory(GT).The basic idea of reliability generalization is that,on the bases of acknowledging the reliability co-efficient in testmaking,emphasized on analyzing the test's reliability co-efficiency and its variability.By using statistical analysis,researchers try to find the source of variability,and furthermore,show more appropriate sample community attributes and characteristics for the test.This technique can assist researchers to acquire valid and reliable scores without any mistake of reliability introduction.Thus,it results to make more accurate and effective test in order to increase its scientific reliability.
出处
《学术研究》
CSSCI
北大核心
2009年第2期54-59,共6页
Academic Research
基金
广东省普通高校人文社会科学“十五”规划研究项目“现代计量心理学的理论与应用研究”(065DXMXLX002)的阶段性成果
