中文摘要：

表征能力发展部分地体现为对事物中隐含关系的理解或推理的复杂程度。基于辛自强的“关系-表征复杂性模型”。结合Halford的“关系复杂性”理论，通过事前分析任务复杂性确定了难度不同的项目，最终编制成儿童图形表征能力测验，该测验共50个（25类）项目。以265名小学3～6年级儿童为被试，采用计算机呈现方式施测了该测验。结果表明：项目的难度和区分度以及整个测验的重测信度、构想效度、效标关联效度均良好；基于理论的事前任务分析（确定任务的维度数与每个维度内的元素数）对事后任务难度的预测率较高（83．4％）。可见，所编制的图形测验能有效测查儿童表征水平。

英文摘要：

Representational capacity can be measured in terms of the complex degree of relations individual can comprehend and understand. Based on the relational- representational complexity model developed by Xin Ziqiang, combined with the relational complexity theory proposed by Halford, a computerized Figuml Representational Capacity Test was designed for measuring children. The test consisted of 50 items, which belonged to 25 categories. The test was used in the study to measure 265 subjects from grade 3, 4, 5 and 6. Results showed that： （1） Most items＇ difficulty, distinctiveness were good; （2） Reliability, constructs validity, criterion related validity of the test were also satisfactory; （3） The prior analysis of task complexity （involving analysis of dimensions in every item and elements included in every dimension） based on the two theories could fully predict task difficulty （ R^2 = 0.83）. Such results indicted that the test could be an effective and reliable instrument for measuring children＇s representational level.