Poster #115 - The role of dialect in African American children’s arithmetic performances: Insights from error patterns

• In Event: 3-206 - Poster Session 13
  In Poster Session: PS 13 Section - Language, Communication

Integrative Statement

Though arithmetic tasks are traditionally measured with frequencies of correct/incorrect responses, a growing body of research has indicated that children’s arithmetic strategies and patterns of error propagation can be indicative of their mathematical development (Mazzocco, Murphy, Brown, Rinne, & Herold, 2013; Ryan & Williams, 2007; Siegler & Shrager, 1984). These qualitative metrics of mathematical performance depend on direct contributions of both child abilities (e.g., language proficiency) and item features (e.g., problem-size; Campbell & Epp, 2005; Geary & Wiley, 1991; Imbo & Vandierendonck, 2007; LeFevre, Sadesky, & Bisanz, 1996; Siegler, 1991; Siegler & Shrager, 1984; Siegler & Taraban, 1986). Unfortunately, African American children, many of whom speak a non-mainstream dialect of American English are under-represented in this body of research. Thus, their status as a cultural and linguistic minority has remained largely unexamined as a potentially important predictor of problem-solving. The purpose of this study was to describe the relationship between African American children’s proficiency with mainstream American English dialect (MAE) and their qualitative patterns of performance across various formats of arithmetic tasks.

This study used a sample of 42 (n = 25 male), typically-developing African American 2nd and 3rd grade students, living in a large metropolis in the Southeastern United States, all of whom exhibited strong dialect variation from MAE. Participants were administered the Diagnostic Evaluation of Language Variation Screener (DELV-S; Seymour, Roeper, & de Villiers, 2003) Language Variation subtest as well as an experimental math assessment, which is described in Table 1.

A series of multilevel logistic regression models (PROC GLIMMIX) were fit to account for item and person features. (SAS Institute Inc., 2013). The best-fitting model results indicated that children with higher MAE dialect proficiency were more likely to correctly answer math items regardless of item features (B = .18, SD = .07, p < .01). However, children’s proficiency with MAE significantly interacted with item-language formatting to predict the likelihood of selecting an inappropriate strategy for solving problems (see Table 2). On language-formatted items, each unit increase in MAE proficiency was associated with a -.25 logit decrease in the likelihood of selecting an inappropriate strategy (p < .01).

Taken together, results from the current study suggest that African American children may have difficulty forming accurate mental representations of what word problems are asking them to do as a function of their proficiency with MAE. A traditional analysis of correct/incorrect answers would not generally reveal these difficulties. Rather, an examination of the quality of children’s errors was necessary to observe this effect. A growing body of research has begun to explore the possibility that students who are cultural and linguistic minorities may encounter unique difficulties with mathematical achievement as a result of hidden cultural-linguistic problem demands (Abedi & Lord, 2001; Martiniello, 2009; Rhodes, Branum-Martin, Morris, Romski, & Sevcik, 2015; Shaftel, Belton-Kocher, Glasnapp, & Poggio, 2006). These results lend support to the idea that language-based mathematics assessments, which implicitly rely on skill with MAE dialect, may evidence cultural bias for AAE users as a function of their proficiency with MAE.

Authors

• Katherine Rhodes, The Ohio State University
  Presenting Author

• Julie Washington, Georgia State University
  Non-Presenting Author

• Lee Branum-Martin, Georgia State University
  Non-Presenting Author

• Stephen Petrill, The Ohio State University
  Non-Presenting Author