General Vocabulary, Academic Vocabulary, and Vocabulary Depth: Examining Predictors of Adolescent Reading Comprehension

• In Event: Roundtable Session 60
  In Roundtable Session: Depth and Breadth: Approaches to Assessing Partial Vocabulary Knowledge and Supporting Word Learning

Abstract

Goals/Purposes
To examine (1) relationships among general and academic vocabulary and reading comprehension for middle-grade students, and (2) relationships between several measures of academic vocabulary and reading comprehension for students in middle grades.

Perspectives/Theoretical Framework
Young children learn words rapidly through early childhood (Anglin, 1994) and there is some evidence that the order of acquisition is consistent across children through early elementary grades (Biemiller & Slonim, 2001). Older children continue to learn new words, but these are often low-frequency words that they tend to encounter through reading rather than through aural exposure (Nagy, Anderson, & Herman, 1987; Nagy, Herman, & Anderson, 1985). Within the group of relatively low-frequency words that might be considered “academic” vocabulary, there are some that are ubiquitous across disciplines (such as process, alternatively, and therefore) and others that are found within specific domains (such as photosynthesis or acute angle). Many researchers have argued that cross-content academic vocabulary has a strong relationship with reading comprehension and therefore deserves additional instructional time (Beck, McKeown, & Kucan, 2002; Lawrence, White, & Snow, 2010). To date, there have been few studies that explore the impact of instruction on general academic words on student reading comprehension (Snow, Lawrence, & White, 2009) and none that we know of that explore these issues with standardized reading measures or sophisticated vocabulary depth measures.

Methods/Techniques
A preliminary analysis using multivariate regression has been completed. Equating/linking and IRT for the common anchor set and items shared with related studies (see the first paper in this session) and trend analysis with the thetas derived from this linking is underway.

Data Sources/Evidence.
2,615 6th, 7th, and 8th graders in urban school districts participated in a vocabulary intervention organized in 2009-2010. A vocabulary assessment was administered as pre- and posttest, containing 50 multiple-choice synonym vocabulary items, used as a common anchor, followed by one of twenty forms comprised of 10-12 homogenous groups of depth items. There were three types, each intended to measure different aspects of partial vocabulary knowledge: idiomatic (collocational patterns), topical (word associations), and hypernym (semantic class). Test forms were spiraled and randomly distributed among students. Each student was also administered the Gates-MacGinitie Reading Test.

Results

There were high correlations between measures of academic and general vocabulary (r = 0.733, p < 0.001), and also between general vocabulary and reading comprehension (r = 0.726, p < 0.001) and academic vocabulary and reading comprehension (r = 0.703, p < 0.001). In a series of forced hierarchical regressions predicting reading comprehension we found that academic vocabulary knowledge explained an additional 6.2% of variance in reading scores above and beyond that explained by the Gates global vocabulary assessment. Furthermore, inclusion of depth measures for the high leverage academic words were significant in subsequent models, and their inclusion improved the regression models. These analyses will be improved using thetas derived from the within-subject study described above; however, they suggest that knowledge of the narrow band of cross-discipline vocabulary words has a relationship with reading comprehension above and beyond general vocabulary knowledge.

Authors

• Joshua Fahey Lawrence, University of California

• Elizabeth Juliana Pare-Blagoev, Harvard University

• Rene R. Lawless, Educational Testing Service

• Paul Deane, Educational Testing Service

• Chen Li, ETS