【作者】 张龙飞； 刘凯； 宋鸽； 蔡艳； 涂冬波；

【Author】 Longfei Zhang;Kai Liu;Ge Song;Yan Cai;Dongbo Tu;School of Psychology,Jiangxi Normal University;

【机构】 江西师范大学心理学院；

【摘要】 情绪智力是当前心理学的研究热点之一,情绪智力能够解释以下现象:有些在传统意义上智力较低的个体却常常能够取得比其他传统智力较高的个体更高的学业和工作成就。以往研究发现在对于个体成就的预测上,情绪智力能够比传统意义上的智力表现得更好。现阶段对于情绪智力的测量主要通过经典测量理论(CTT)框架下的纸笔测验来实施,纸笔测验的耗时较长,测量误差较大,且CTT作为传统的测量理论具有许多固有的缺点。而本研究建立在新一代的测量学理论——项目反应理论(IRT)框架下,首次尝试采用计算机化自适应测验(CAT)技术实现对情绪智力的智能测评。计算机化自适应测验技术可以根据被试当前作答情况,实时计算并更新其特质水平,并把能提供最大信息量的项目匹配给被试进行作答,以此获得耗时短、精度高的测量效果。本研究首先基于IRT系列分析(含单维性检验、模型拟合检验、局部独立性检验以及项目质量分析),构建了符合IRT测量学要求的情绪智力测评的题库,并以此为基础探讨了计算机化自适应测验技术在情绪智力智能测评(CAT-EI)中的应用。结果表明:(1)CAT-EI相关算法具有较高的参数估计精度,同时具有较理想的测量信度和效度;(2)CAT-EI可以使用较少的题量(Mean=9.88题)达到使用整个题库67题的测量精度,它一方面能做到减轻被试的测试负担,另一方面实现了对情绪智力的高效、快速、准确的智能测评。总而言之,计算机化自适应测验技术可以克服传统纸笔测验的缺陷,通过使用该项技术,本研究能够开发一种新型的、基于计算机的、自适应的情绪智力智能测评。更多还原

【Abstract】 Emotional intelligence maintains an important position in current research of psychology due to its effective prediction and influence on daily lives. The measurement of emotional intelligence in today relies on pen-pencil test, which is time-consuming, and will cause high measuring error. These shortcomings comes from pen-pencil test can be solved easily by the computerized adaptive testing(CAT). CAT is a form of testing that uses a computer to automatically select appropriate items for the examinee(Almond & Mislevy, 1998). It can largely reduce respondent burden that means saving considerable time for the examinees and improve the measuring precision by the means of selects items which is most suitable to the examinee’s theta from an item bank. Therefore, developing a computerized adaptive testing for the measurement of emotional intelligence is a useful, and urgent meaningful task in the filed of psychometrics.This study aimed to examine the psychometric properties of the item bank of emotional intelligence and develop a computerized adaptive testing for emotional intelligence(CAT-EI) to assess emotional intelligence characteristics in the framework of item response theory(IRT). 865 participants who were all volunteer for this study and whom we obtain informed consent in accordance with the Declaration of Helsinki from was recruited. Results show that: 1) The final item bank of emotional intelligence with strict requirements of IRT contained 67 items which owned high item information. And the mean of IRT discrimination of the item bank reached 1.22, which showed that the item bank of CAT-EI was high quality; 2) CAT-EI under the stop rule of SE<0.4 made an good estimation of characteristics(the correlation coefficient between CAT-EI and the item bank of emotional intelligence is 0.92) and had acceptable and reasonable marginal reliability(MR=0.85), criterion-related validity(the correlation coefficient is 0.79 with REIS-C, 0.76 with EIS-C and 0.87 with WLEIS-C); 3) CAT-EI reached the measurement precision of the whole CAT-EI bank(67 items) with the average of 9.88 items under the stop rule of SE<0.4. There are also some directions that should be addressed in future: 1) The item selection algorithm of CAT-EI is based on the maximum Fisher information(MIF) criterion which means that it will choose the items whose discrimination were high more often and this kind of items risk being exposure to the examinees; 2) The size of item bank of emotional intelligence was small for such computerized adaptive testing. And it needs to be extended by adding more good items for emotional intelligence measuring. In a whole, this study tried to employ computerized adaptive testing to assess emotional intelligence(called CAT-EI). It not only can reduce respondent burden for the examinees that means saving a lot of time for them, but also can make an efficient, quick and accurate assessment for emotional intelligence. In brief, a new psychometric technology of emotional intelligence was provided by this study and can use this kind of technology to replace the place of pen-pencil test of emotional intelligence commendably.更多还原