This dissertation consists of two manuscripts reporting on patterns in and predictors of fourth-grade students' reading performance. By analyzing the data from the Progress in International Reading Literacy Study (PIRLS), the studies in this dissertation focused on student-level and classroom-level information in order to explore significant predictors of U.S. students' reading performance. The multilevel model was the main analytical approach of the studies in this dissertation, which involved students nested within classrooms. The studies drew on a complex view of the development of reading.

In the first study of this dissertation, I explored predictors of student performance on two different purposes of reading, that is, informational reading and literary reading. The inspiration for this study came from the fact that U.S. students have a large performance gap favoring literary reading over informational reading on average according to the reports from PIRES 2001 and 2006. In order to understand this gap, the study examined differential contributions of literacy-related variables to informational versus literary reading performance on a country-level and within selected countries including the United States. The results indicated that several genre-related activities and curriculums were significantly related to fourth-grade children's reading performance on either literary or informational reading and, in some cases, across both of these genres. In the U.S., for instance, more instruction with informational text in classroom predicted higher reading performance in informational reading. The implication of this study is that substantial opportunities to learn with both literary and informational text and also opportunities to learn text style and structure should be provided in primary grades for all children.

The second study focused on the variables of reading motivation, based on their strong relationships to student reading performance shown in the first study. The purposes of the study were to explore underlying factors in reading motivation measured in PIRLS and to investigate the relationships between those factors and reading performance. Exploratory and confirmatory factor analyses revealed that motivational items in PIRLS consisted of four different factors which also could be clustered as two contrasting facets (internal orientation and external orientation) under two different dimensions (attitudes toward reading and reading self-concept). Further analysis discovered interaction effects between internally oriented motivation and externally orientated motivation to predict better reading performance. Although the relationship of one motivational facet to reading performance was different depending on the levels of its contrasting facet, internally oriented motivation was always a positive predictor of student reading performance. Extrinsic motivation under the dimension of attitudes toward reading, for example, positively related to reading performance if a student had high intrinsic motivation. However, it was a negative predictor of reading performance for those students with low intrinsic motivation. Finally, four motivational factors together explained 17% of the student-level variance in overall reading performance.