'If I write like a scientist, then soy un cientifico': Differentiated Writing Supports and the Effects on Fourth-Grade English Proficient Students' and English Language Learners' Science Content Knowledge and Explanatory Writing About Magnetism and Elect

The purpose of this pre-post quasi-experimental dissertation was to investigate the effects of differentiated writing supports on English Proficient Students' (EPSs) and English Language Learners' (ELLs) science content knowledge and explanatory writing about magnetism and electricity. Eighty-seven fourth-grade students (EPSs = 35; ELLs = 52) were randomly assigned to two groups based on two differentiated writing: guided questions ( n = 43) or targeted writing frames (n = 44). In the guided questions condition, students completed four question sets after a science investigation, and in the targeted writing frames condition, students completed the same four question sets, but with explicit support for vocabulary, transitions, and relational language in the form of if-then statements. Over the course of the four week intervention, students completed a total of nine writing tasks, and were pretested and posttested on six variables: magnetism and electricity content knowledge test, explanatory writing task, total number of words written, total number of sentences written, number of if-then statements, and number of content-based vocabulary words. Results indicate that EPSs and ELLs in both writing conditions improved significantly from pretest to posttest on six content and explanatory writing variables, with statistically significant gain scores occurring for the magnetism and electricity content knowledge test in which the targeted writing frames condition had a larger rate of gain. ANCOVA results indicated that in comparing writing conditions, a statistically significant difference was found for magnetism and electricity content knowledge posttests, when controlling for pretests. No statistically significant effects for language classification on the six variables were found when controlling for pretest scores. Interaction effects between writing condition and language classification were statistically significantly different for the interaction effect found on if-then statements in which ELLs in the writing frames group included the most if-then statements. Implications of these results and future research possibilities are addressed.