Title

Toward Creating Computer-Based Math Learning Favoring High-School Females

Document Type

Presentation

Journal/Book Title/Conference

The Annual Conference of Women in Engineering Proactive Network (WEPAN)

Publication Date

6-17-2009

Abstract

Research indicates that teenage females prefer to work and perform better at the learning environment that supports frequent interactions and allows them to build relationships with others. This paper will introduce a computer-based algebra-learning environment MathGirls equipped with pedagogical agents (digital life-like characters) that simulate real-world social interactions and relations. The goal of MathGirls is to help young women of high-school age build positive attitudes toward and self-efficacy in math learning through this simulated social context. To investigate the efficacy of MathGirls, a classroom experiment was conducted with 83 high-school females. The experiment examined the effects of agent attributes (female teacher-like, male teacher-like, female peer-like, and male peer-like) and learner ethnicity(Caucasian vs. minorities). The results indicated that minority females’ math attitudes were significantly higher after working with the female-teacher-like agent than with the other agents. Second, Caucasian females’ math attitudes were significantly higher after working with the male agents, whereas minority females’ math attitudes were significantly higher after working with the female agents. Third, both groups of the female students perceived the peer-like agents to be significantly more affable than the teacher-like agents. Fourth, Caucasian females performed significantly higher after working with the teacher-like agents, whereas minority females performed higher with the peer-like agents.From the results, the author conjectures that the concept of attribute similarity (the similarity in personal characteristics of a role model and a learner) seems to have applied more to minority females. In contrast, the projection of real-world social stereotypes seems to have applied more to Caucasian females. Overall, the findings support the instructional utility of agent attributes to address the needs of diverse groups of young female learners.

This document is currently not available here.

Share

COinS