There is an established gender gap in middle school math education in the U.S., where girls report higher anxiety and lower engagement than boys, which negatively impacts their performance and even long-term career choices. While digital learning games, with the ability to promote learning motivation and outcomes, have the potential to address this gap, there have been mixed results regarding gender differences in learning with games. Furthermore, prior research on the gender effects of learning games remains focused on the distinctions between boys and girls, without accounting for the spectrum of variance in gendered behavior, which can develop as early as middle school.

In my work, I have identified Decimal Point, a digital learning game that teaches decimal numbers and operations to middle school students, as an excellent platform for studying gender effects in digital learning games. Based on data from five prior Decimal Point studies with over 1,000 students, I have observed a consistent gender difference across all studies &nash; girls tended to have lower prior knowledge, but better self-explanation performance and higher learning gains from the game. The difference in self-explanation also explains the relationship between gender and learning outcomes, suggesting that girls' better learning could be attributed to their self-explanation performance. At the same time, there were no differences in how boys and girls enjoyed the game, indicating that digital games can help close the gap in learning while not sacrificing enjoyment for all students.

I conducted two follow-up studies to better understand the reasons for the observed gender differences and the extent to which they generalize. In both studies, I also employed a multidimensional representation of gender, one that captures not only birth-assigned gender and gender identity, but also gender-typed occupational interests, activities and traits. The first study investigated how self-explanation and different learning platforms influenced the relationship between gender and learning outcomes. The second study examined the ways in which gender differences in learning outcomes and enjoyment manifested when changing the game narrative. Results from both studies indicated that, across different learning platforms and game narratives, girls learned more than boys thanks to their better performance in the self-explanation activities. Furthermore, analyses of multiple gender dimensions led to a more nuanced understanding of the gender effects than analyses of binary gender alone. For instance, while boys generally reported higher levels of engagement with the game than girls, analyses of multidimensional gender further revealed that girls with strong masculine-typed behaviors were also more engaged than others.

In summary, this work contributes (1) robust evidence of the benefits of self-explanation in helping girls learn from digital games, (2) insights on the use of multidimensional gender representation to capture nuances in gender differences with respect to learning, enjoyment and game preferences, (3) guidelines for designing effective learning games to bridge the gender gap in math education. In a broader sense, this research will advance knowledge on the multidimensionality of gender in learning game research and inform practical recommendations on aligning game features with individual learners to optimize learning and engagement.

128 pages

Thesis Committee:
Bruce M. McLaren (Chair)
John Stamper (Chair)
Jodi Forlizzi
Jessica Hammer
Derek Lomas (Delft University of Technology)

Brad A. Myers, Head, Human-Computer Interaction Institute
Martial Hebert, Dean, School of Computer Science

Return to: SCS Technical Report Collection
School of Computer Science homepage

This page maintained by reports@cs.cmu.edu