Expanding Knowledge of Connectives in Spanish/English Emergent Bilinguals in Second Grade

• In Event: 31.088 - Deconstructing Vocabulary to Inform Practice: Investigating Different Aspects of Vocabulary Knowledge With Diverse Learners

Abstract

Objective

The purpose of this study was to design, implement, and analyze effects of a ten-week intervention to teach school-relevant connectives (shown in prior work as difficult to acquire by third graders) to Spanish/English emergent bilinguals in second grade.

Theoretical framework

Knowledge of connectives—words and phrases that signal logical relations between ideas, e.g., meanwhile, because—is an important contributor to reading comprehension (e.g., Degand & Sanders, 2002; Graesser & Millis, 1997). Spanish/English emerging bilinguals have been found to lag behind their native English-speaking peers in connective knowledge (Crosson & Lesaux, 2013).

Methods

Driven by pedagogical need to develop high-utility school-relevant language early on, this intervention was conceptualized to target second graders rather than older children (Snow, Griffin, & Burns, 2005).

Connective knowledge was first explored with 388 third graders in the Southwest and Northeast and those findings informed the word selection for the intervention. From those difficult to acquire connectives, twelve were selected for intervention using the following criteria: first, to target the most relevant connectives, connectives with high frequency in primary and later elementary grades were selected (Zeno, Millard, Ivenz, & Duuvari, 1995); second, connectives were stratified by word level according to the Living Word Vocabulary (LWV, Dale & O’Rourke, 1981) and logical relation signaled (i.e., aspectual, temporal, causal, adversative, contrastive, and conditional) to explore possible moderating effects.

Using a randomized control design, 22 transitional bilingual classrooms were randomly assigned to treatment (full intervention) or control (materials only) within eight Southwestern schools. A pedagogical approach was designed to provide authentic communicative purposes, engage students in text-based conversations, represent and express connectives through multiple modalities, playfully engage students in learning, encourage language production, and gradually reduce scaffolding (Graves, August, & Mancilla-Martinez, 2013; Uccelli, Rosenthal, & Barr, 2011).

Data sources

Third grade participants in the first study were assessed with a researcher-developed measure modified from a task reported in Crosson, Lesaux, and Martiniello (2008). For the intervention, second grade participants were pretested in English on the Picture Vocabulary, Syntactic Understanding, and Sentence Imitation subtests of the Test of Language Development (TOLD), as well as a curriculum-based assessment of the twelve selected connectives. Nesting of subjects within classrooms was accounted for in hierarchical linear regression analyses by specifying multi-level models with two discourse-level assessment outcomes: a cloze assessment measuring participant’s ability to select correct connectives to complete sentences about a children’s story, and a written retelling measuring participant’s production of connectives while retelling a children’s story.

Results

Controlling for pre-test knowledge of connectives, preliminary findings indicate a positive effect of the intervention on post-test knowledge of connectives on two researcher-created discourse-level assessments with differential effects for different levels of language knowledge at pretest.

Significance

Findings have timely implications for practice considering growing commitment by school districts nationwide to the Common Core Standards that requires primary grade students to understand and use connectives early on. This study contributes to the research literature, expanding our understanding regarding connectives and how engaging interventions might positively impact students’ connective knowledge even in early primary grades.

Authors

• Lauren Artzi, American Institutes for Research

• Lindsey Anne Massoud, Center for Applied Linguistics

• Franne Rosenthal, Harvard University

• Paola Uccelli, Harvard University

• Christopher Daniel Barr, University of Houston

• Diane L. August, American Institutes for Research