LAUSR

We examined the contributions of English proficiency, genre, and the use of textual sources to adolescent writing. The sample included 1819 native English speakers and language minority students from 127 seventh- and eighth-grade classes in an urban school district. Students were randomly assigned one of three source-based essay prompts (narrative, explanatory or argumentative) as part of the annual state assessment, and all students wrote a common, on-demand argumentative essay that did not require the use of textual sources. Overall, language minority students who were fluent English proficient wrote higher quality nonsource-based essays that contained more difficult vocabulary, were better structured, and used examples and details more effectively than native English speakers. They also outperformed native English speakers in source-based writing. Students with limited English proficiency showed weaker performance on both source-based and nonsource-based writing tasks. Differences in English proficiency held across genres. Students who wrote argumentative essays obtained higher ratings than those who wrote narrative or explanatory essays. Source-based and nonsource-based writing were moderately correlated. Regression analyses revealed that in addition to English proficiency and genre, reading comprehension's contribution to source-based writing was almost double that of nonsource-based writing. Implications of the findings for theory and practice are discussed.