Words in Motion: An intervention to promote the academic vocabulary development of adolescent native Spanish-speakers

• In Event: The Development of Vocabulary and Comprehension in Spanish-Speaking English-Language Learners

Abstract

Objectives
It is well established that vocabulary is a critical component of literacy development and that many children, particularly second language learners, lack the academic vocabulary necessary to comprehend school texts (August & Shanahan, 2006). One promising approach for developing the vocabulary of native Spanish speakers in particular is to focus on cognates (August, Carlo, Dressler, & Snow, 2005). This presentation addresses the following question: Relative to each other and to no treatment, what are the effects of a cross-linguistic intervention and a monolingual English intervention on the academic vocabulary skills of middle school students?

Theoretical Framework
The intervention is based on three theoretical pillars: Michael Graves’ 4-part framework for vocabulary instruction (Graves, 2006), sheltered instruction (Echevarria, Vogt, and Short, 2004), and responsive pedagogy (Gay, 2000).

Method
The pilot phase of this study involved 346 6th, 7th, and 8th grade students in a single urban middle school with a predominantly Latino, native Spanish-speaking population. Six classrooms received the monolingual English curriculum (MEC), six classrooms received the cross-linguistic curriculum (CLC), and nine control classrooms received the standard language arts curriculum. The curriculum consisted of five, 10-word units taught over 31 non-continuous days. Because of scheduling constraints, 8th graders received only the first three units.

Data Sources
A study-specific vocabulary measure was administered as a pre-test, as a post-test for all students at the end of unit 3 (post 1), and again to only 6th and 7th graders at the end of unit 5 (post 2). The EVoCA vocabulary test is a 74-item multiple choice assessment (α=.90) that includes taught cognates, non-taught cognates, and non-taught non-cognates.

Results
Pre/Post 1 and Pre/Post 2 mean comparisons for all items on the test yield comparable effect sizes for the control group (.29, .29) and the CLC group (.20, .26), while effect sizes for the MEC group are significantly larger (.44, .66). However, when conducting mean comparisons of only the taught items, effect sizes for the CLC group are also significantly larger than those for the control group (.22 at both p1 and p2). At post 1, the effect sizes for the CLC group (.62) and MEC group (.70) are similar, whereas at post 2, the effect size for the MEC group is considerably larger (.91 vs. .58 CLC). The consistent advantage of the MEC condition is surprising given the high density of Spanish-speakers in the school and the availability of bilingual education in the district. Project researchers have recently completed data collection on the final intervention at full scale (4 schools) and will replicate these analyses with the new dataset to see if they are consistent.

Scientific Significance
This intervention study extends the work of previous cognate studies by comparing the effects of the intervention delivered solely in English as a general academic vocabulary curriculum vs. delivered through English and Spanish as a cognate curriculum, with explicit connections made across languages.

Authors

• Elizabeth R. Howard, University of Connecticut

• Igone Arteagoitia, Center for Applied Linguistics

• Eileen Mercedes González, University of Connecticut

• Jennifer D. Green, Western Washington University