| The present study investigated longitudinal changes of the reading achievement among schools populated with English learners. It also examined the heterogeneity in the English learners group in terms of students' performance in high stakes reading tests. Historically, English learners have often been considered the students who are in the process of developing English language competence. These students represent one of the fastest-growing groups among the school-aged population in the United States. However, there have been significant achievement gaps between English learners and native English-speaking students in all grades and content areas. The first set of research questions provided empirical evidence for whether the reading performance of schools with English learners changed over time and how the changes varied across schools. The second set of research questions in this study confirmed that there exist unobserved, multiple classes within this group with background covariates that influence the classification and indicators of reading achievement scores.;The current study employed a secondary data analysis using an existing dataset collected from a large urban school district in the southwestern United States. In the database, a total of 13,060 students enrolled from 2000-01 to 2003-04 in ten schools. The standardized reading achievement scores and background/ language variables of these students were examined. A three-level Hierarchical Linear Model (3-Level HLM) and a Latent Class Analysis (LCA) model were used to analyze the dataset. Findings from the first set of research questions revealed that there were significant growth factors in schools' initial status of reading achievement but not in their rate of change. The effects of background variables such as SES, ethnicity, and language status were statistically significant. Limited English proficient (LEP) students showed significantly lower achievement, controlling for other background variables. Second language oral language proficiency was found to be a major predictor in reading for English learners. In addition, significant school-level effects were found in the school reading proficiency and the percentage of teachers who provided bilingual instruction. Results from the Latent Class Analysis (LCA) demonstrated that there were three unobserved subgroups that consisted of a mixture of distributions among a sample of English learners. Three groups were found to be qualitatively different as they presented low, middle, and high performances in the latent indicators. Covariates such as gender, SES, birth country and grade began in the U.S. significantly contributed to classification and they also predicted reading outcomes. The implications of these findings were discussed in terms of reading research and practice of English learners, and suggestions for future research were presented. |
