阅读是个体获取知识的重要途径，阅读能力的高低关乎个人未来的长期发展和终身学习的能力；阅读投入（reading engagement）与阅读能力的发展密切相关。阅读投入的概念已在西方文化背景下展开了大量研究，并取得了一定的研究成果。然而，在我国教育和科研背景下，还甚少有研究者专门针对这一概念展开研究。了解阅读投入的结构、影响因素及其与阅读成绩的关系，有助于更好地培养青少年儿童的阅读投入，改善他们的阅读态度和动机，形成良好的阅读习惯，为未来的工作和生活奠定坚实的基础。针对以上问题，本研究以2009年学生能力国际评价（PISA 2009）项目的中国大陆预实验数据和美国数据为研究对象，其中，中国学生18446名，学校540所，美国学生5233名，学校165所；依托PISA 2009的学生问卷和阅读素养测试为工具，采用验证性因子分析、多水平线性模型和多层级中介效应分析方法，考察了阅读投入的结构、影响阅读投入的因素以及阅读投入分别在各学生与学校水平自变量对阅读素养影响的中介效应是否显著，及其在中美两国学生之间的表现是否具有跨文化的一致性。整体来看，本研究主要取得了以下发现： （1）中、美两国学生具有相同的阅读投入结构，且该结构具有跨文化的一致性，即阅读投入的一阶8因子模型——阅读时间、工具性阅读、愉悦性阅读、阅读材料的多样性、网上阅读活动的多样性、记忆策略、控制策略、精细化加工策略。 （2）在中美合并的全样本数据中，部分验证了两类更加简约的二阶模型：二阶3因子模型（情感投入、行为投入和认知投入）和二阶单因子模型（阅读投入）；但是在中国和美国的单样本数据中，却拒绝了这两类二阶因子模型的假设。 （3）尽管中美两国学生在阅读投入的结构维度上一致，但是两国在一阶因子与观测变量间的载荷却有显著性差异；即中美两国在阅读投入结构上的形态模型相同，但并不支持因子载荷等同模型。 （4）中国学生在自我报告的阅读时间、工具性阅读、愉悦性阅读、精细化加工策略四个维度上的均值显著高于美国，而美国学生则在自我报告的阅读材料的多样性、网上阅读活动的多样性、记忆策略、控制策略四个维度上的均值显著高于中国。然而，根据Cohen’d效应值的标准判断，可以发现中美两国学生在阅读投入上的均值差异仅在工具性阅读、网上阅读活动的多样性和控制策略三个维度上的效应值处于中等或较大的水平，其他五个维度上中美均值差异的效应值均小于0.40，属于较小的水平。 （5）中国学生除在阅读时间和阅读材料的多样性维度上无显著的性别差异外，其他6个维度上均存在显著的性别差异，美国学生则在8个阅读投入维度上均存在显著的性别差异。父母最高的职业地位在中美两国学生阅读投入8个维度上的回归系数值均非常小且绝大部分均不显著。父母最高的受教育水平和家庭教育资源对阅读投入8个维度的作用模式不相同，在中美之间的作用模式也不一致。文化拥有物和两个课堂教学因素（教师对阅读投入的激励和结构化辅助化策略的使用）对中国学生的工具性阅读具有显著的正向预测作用，但对美国学生的工具性阅读则具有显著的负向预测作用，总体上来看，这三个变量均能显著促进阅读投入的其他7个维度的发展，并在中美两国表现一致。学校因素中的学校氛围和师生关系对阅读投入8个维度的作用模式不相同，在中美之间的作用模式也不一致。 （6）阅读投入对阅读素养多层级中介效应的中美比较分析显示，阅读投入中介效应显著的维度在中美之间并不一致，显著的中介效应的作用模式在中美之间也并不完全一致，甚至出现完全相反的情况。其中，对中国学生来说，工具性阅读（a和b均为正值）、网上阅读活动的多样性（a为正值，b为负值）、控制策略（a和b均为正值）、精细化加工策略（a和b均为正值）在各自变量对中国学生阅读素养影响的中介效应均显著，其他4个阅读投入维度的中介效应不显著；对美国学生来说，工具性阅读（a和b均为负值）、愉悦性阅读（a和b均为正值）、网上阅读活动的多样性（a和b均为正值）、控制策略（a和b均为正值）、精细化加工策略（a为正值，b为负值）在各自变量对美国学生阅读素养影响的中介效应均显著，其他3个阅读投入维度的中介效应不显著。

Reading is an important way to achieve knowledge. Reading ability is critical to the long-term development and life-long learning ability. There have been many studies about the concept of reading engagement under the western culture and accomplished some findings. However, there are few researchers to do this field in China. Learning the structure and influencing factors of reading engagement and how it works with reading achievement are good for cultivating chldren’s reading engagement, improving their reading attitude and motivation, developing favorable reading habits and providing solid foundation for the future work and life. Based on the data of Mainland-China and USA of PISA 2009 (there were 18446 students and 540 schools for China, 5233 students and 165 schools for USA), this study employed the PISA 2009 Student Questionnaire and Reading Literacy Test as the instruments and investigated the structure and influencing factors of reading engagement and its relationship with reading literacy. The data analysis methods included confirmatory factor analysis (CFA), multilevel linear model and multilevel mediation model method. The findings indicated that: (1) There was the same structure of reading engagement for both Chinese and American students. That is, the first-order 8 factors model of reading engagemnt --- reading time, instrumental reading, enjoyment reading, variety of reading materials, variety of online reading, memory strategies, control strategies, and elabration strategies. (2) Two simpler second-order models, the second-order model with 3 factors (emotinal engagement, behaviroal engagement, and cognitive engagement) and the secondorder model with single factor (reading engagement), were partly verified by the full data sample with both China and USA data. However, these two models were not be validated for China and USA data separately. (3) Although China and USA shared the same structure of reading engagement, there were significant differences on the factor loadings between China and USA. In other words, China and USA had the same configural model, but didn’t share the same factor loadings. (4) The mean of each four reading engagment --- reading time, instrumental reading, enjoyment reading, and elabration strategies for Chinese students was significantly higher that each that of USA students, while the mean of each another four reading engagment --- variety of reading materials, variety of online reading, memory strategies, and control strategies for USA students was significantly higher that each that of Chinese students. Nevertheless, according to the criteria for the effect size of Cohen’s d, only the mean differences of instrumental reading, variety of online reading, and control strategies for China and USA were medium or large, the mean differences of another five dimensions were small. (5) There were no significant gender differences on reading time and variety of reading materials, but significant gender differences were found on another 6 dimensions for Chinese students. Significant gender differences were all shown on the 8 dimensions for USA students. The regression coefficients of parents’ highest occupational status on the 8 dimensions of reading engagement were all very small and mostly not significant for both China and USA. Parents’ highest educational level and home educational resources had different impacts on the 8 dimensions of reading engagement, and also were different between China and USA. Cultural possessions and two elements about classroom teaching (teachers’ stimulation of reading engagement and use of structuring and scaffolding strategies) had significantly postive effects on instrumental reading for Chinese students, while significantly negtive effects for USA students. As a whole, these three variables had significant postive effects on the other 7 dimensions, which were consistent for the two countries. The two school factors --- school climate and teacher-student relationship had different impacts on the 8 dimensions of reading engagement, and also were different between China and USA. （6）The results of multilevel mediation effect analysis indicated that the dimensions of significant mediation effects were different for China and USA, also with the influence patterns. For Chinese students, insturmental reading (both a and b were positive), variety of online reading (a was positive, b was negtive), control strategies (both a and b were positive), elaboration strategies (both a and b were positive) were separately significantly mediated on the relationship of various independent variables and Chinese students’ reading literacy. However, for USA students, insturmental reading (both a and b were negtive), enjoyment reading (both a and b were positive), variety of online reading (both a and b were positive), control strategies (both a and b were positive), elaboration strategies (a was positive, b was negtive) were separately significantly mediated on the relationship of various independent variables and USA students’ reading literacy.