基于儿童汉语分级阅读的阅读能力与文本难度的分值测算

Level reading aims to match in scales between reading ability and text difficulty, whose precondition is to set up level criterion. On the basis of the hypothesis that reading comprehension is the subtraction of the text difficulty from reading ability, the key issue is to measure reading ability and text difficulty and transform the resulting coefficient into scales. In English-speaking countries, much research achievement in level reading has been made and applied into practice widely, while in China, Chinese text difficulty formula for children leveled Chinese reading hasn't been worked out. It has been discovered in our prophase research that sentence length variable doesn't contribute to Chinese text difficulty as much as to the English, while sentence pattern does. In our research, the following work will be done:text variables will be analyzed in the connection between linguistics and psychology; Chinese text readability formula will be deduced through experience validation and theoretical validation; anchoring point will be set and the transformation formula will be worked out from text difficulty coefficient to scales; the scaled text will be used as contents of norm corpus from which text items will be compiled and used to test children; the resulting item responses will be input into Rasch Model for getting the scales of reading comprehension, reading ability and text difficulty. With the work mentioned above done, the leveling goal will have been realized. This research has a great application prospect, including children' measuring their own reading ability scales by themselves, publishers' scaling books, homes' and schools' interfering children' reading, and Chinese education as a mother tongue and as a foreign language being improved and standardized.

分级阅读是阅读能力和文本难度在级别上的匹配，其前提是分级的标准，基于阅读理解=阅读能力-文本难度的假设，标准的划分需要测算阅读能力和文本难度并通过公式转换为分值。英语国家的分级阅读研究成熟，成果应用广阔，国内还没有针对儿童汉语分级阅读的可读性公式。我们的前期研究发现，"句长"变量对汉语文本的解释力不如对英语文本那么强，而汉语句式更能影响文本可读性。本课题拟勾联文本可读性变量的语言特征和心理学原理；通过经验验证和统计回归检验，推导文本难度公式，计算文本难度系数；确定分值参照分，构想分值转换公式，将难度系数转换为分值；将赋以分值的文本作为编写阅读理解分级测试题的常模语料库；以分级试题测试个体，将阅读反应项目输入罗斯模型，获得阅读理解、阅读能力和文本难度的分值，实现分级的目标。该研究的应用前景是：儿童自测阅读能力分值；出版社标识难度分值；学校和家庭干预儿童阅读；推动和规范我国母语和对外汉语教学。

分级阅读是儿童阅读的书籍和他的阅读能力相符合，就像儿童穿的鞋子尺码应该适合他的脚一样。鞋子大了，走不快，走的累；鞋子小了，走不了，走的痛。太难的书，儿童看不懂；太容易的书，儿童学不到新知识，也不感兴趣。那么怎么判断这本书对儿童是难还是容易呢？.我们假设，阅读理解=书的难度–阅读能力。我们看不到儿童（假设是小明）的阅读能力，我们看到的是他看了什么书，看的是不是投入，（考试中）他做对了几道阅读理解题目，得了多少分，等等。这些他做了什么和做的怎么样等外在的表现，我们称之为阅读理解，借此我们判断小明的阅读能力。换句话说，如果小明不看书，也就没有阅读理解这个过程和结果，不了解小明的人是不知道他的阅读能力的。再进一步说，如果10岁的小明看的是《儒林外史》，而不是《西游记》，我们判断他阅读能力较高的同时也会怀疑他是否看得懂。也就是之，我们通常是通过一个人所阅读的书籍来判断他的阅读能力的，这实实在在的书（文本）的难度就是我们研究的重点。.研究文本难度的角度很多，我们研究的是作者写作的语言特点，例如段落的长度、句子的长度、句子的类型、用词的通俗性、同一个词重复的次数、笔画数量、词的长短（单字词、双字词、多字词）、词性等。这些特点是客观的，不会突然消失，不会因人而异。我们选择了一批中学生做阅读理解测试，分析测试得分和这批文章中的语言特点，挖掘出显著影响学生得分的若干项语言特点。我们还开展了眼动实验验证这些显著的语言特点，发现两类语言特点验证成功。根据我们的研究结论，假设其余语言特点不变的情况下，这两类语言特点可以预测文本难度的94%。由于我们采用了罗斯模型，使得文本的难度和学生阅读能力具有可比性，即学生阅读能力的差异，94%的原因可以归结为这两类语言变量。.如果出版社根据我们的研究结果对书籍进行难度分级，学生、教师和父母对书籍的选择就有了重要的参考标准，教材内容对学生认知特点的符合程度可能更高。