Are there Matthew effects for a science and history content literacy intervention on near transfer measures of vocabulary depth?

Are there Matthew effects for a science and history content literacy intervention on near transfer measures of vocabulary depth?

First Author: Douglas Mosher -- Harvard University
Additional authors/chairs: 
James S. Kim
Keywords: Vocabulary, Literacy, Intervention, semantic learning, English Language Learners (ELL)
Abstract / Summary: 

Purpose:
The study examined whether an elementary grade content literacy intervention focusing on domain specific vocabulary led to Matthew effects according to students’ baseline characteristics. We investigated the following questions:
1. Are there Matthew effects for ELL students and varying levels of English proficiency?
2. Is there evidence that middle and higher socioeconomic groups benefit from the intervention more than the lowest SES group?
3. Is there evidence that higher performing readers benefit from the intervention more than lower performing readers?

Method:
The intervention and control groups (n=2,142) were recruited from 30 schools in an urban district in the Southeastern US. The vocabulary component of the intervention consisted of teaching semantically related and domain-specific vocabulary using concept maps. We assessed students on both words that were explicitly taught and near transfer words that were not taught but conceptually related to the lesson content.

Results:
Using multilevel modeling, we found that there were no differential treatment effects for ELLs and non-ELLs (p = 0.35) and students with varying levels of English proficiency (p = 0.67). Likewise, there is no indication that the intervention has led to a widening gap for vocabulary acquisition between low and medium SES groups (p = 0.21) or between low and high SES groups (p = 0.97). Finally, there is no indication of Matthew effects for students of varying reading ability (p = 0.71).

Conclusion:
This content-based literacy intervention appears to benefit all students without contributing to Matthew effects. Results show that on average, students were likely able to leverage their conceptual knowledge and make inferences about unfamiliar words that had not been explicitly taught.