Comparing English-only and language-minority learners on English vocabulary knowledge

Language-minority (LM) learners' English vocabularies arc estimated to be much smaller than those of English-only (EO) learners (Grabe, 1991; McLaughlin et al., 2000; Umbel & 0ller, 1994), with discrepancies as large as 90,000 words (Grabe, 1991). The vocabulary knowledge that LM students do possess is also lacking in depth , or quality , in comparison to non-LM learners (August et al., 1999, as cited in August, Carlo, Dressler, & Snow, 2005; McLaughlin et al., 2000; Verhallen 1994, as cited in Schoonen & Verhallen, 2008), even for frequently occurring words (Verhallen & Schoonen, 1993). This poses a problem, given the critical role of vocabulary development in children's literacy development and reading comprehension (Beck, McKeown, & Omanson, 1987; Cunningham & Stanovich, 1997; Muter & Diethelm, 2001; Proctor, Carlo, August, & Snow, 2005; Snow, Burns, & Griffin, 1998). In this dissertation, I report on two studies designed to explore learning differences between LM and EO students, in order to understand vocabulary discrepancies between those groups, and to assess how instructional strategies might be differentiated to better serve LM students' needs. In the first study, I explored whether LM and EO students differed in their vocabulary self-evaluations. I found that English reading achievement, and not language status (LM versus EO), explained the relationship between reported familiarity and English-vocabulary performance. Students who already struggle with their literacy learning, many of whom were LM learners, were less likely to recognize gaps in their vocabulary knowledge. One instructional implication of this work is that teachers need to understand that students' own self-evaluations of vocabulary may be flawed, particularly when they have weak language skills. Future research should investigate whether students can be taught to more accurately self-evaluate their knowledge, and whether this skill facilitates increased learning. For the second study, I analyzed the effect of a universally designed digital reading environment that contained embedded vocabulary instruction and first language (L1) supports on fifth-grade students' understanding of specific words' semantic depth. I found the intervention to be successful for all students. When their English reading achievement was controlled for, LM and EO learners receiving the intervention performed similarly on semantic depth of vocabulary for target words. In the control group, EO learners outscored LM learners on target words, even when controlling for English reading achievement, suggesting that the intervention had an acceleration effect on learning target words for LM learners. Future studies should identify which instructional features are most potent for accelerating the language development of LM students.