Sub-Saharan Africa is the world's fastest growing region with regards to the adoption of mobile technologies with three quarters of the population owning a mobile device¹. The region is also currently undergoing an education crisis with over 32 million out-of-school children, and over 80% of school-enrolled children not meeting minimum proficiency in math and reading assessments. Therefore, interested organizations and governments have attempted to tackle this issue by introducing mobile educational technologies to supplement, and sometimes replace traditional classroom education. Some of these initiatives have shown positive learning gains for students, often attributing these outcomes to children's innate curiosity to teach themselves the content, context-aware curriculums, as well as peer support and collaboration in the learning process. Indeed, peer support should be utilized and encouraged to maximize the efficacy of educational technologies. Collaborative peer learning leads to increased test scores, cognitive activity, motivation and enthusiasm, and satisfaction over individualized learning settings. While knowledgeable adults can provide (or be trained to provide) adequate domain knowledge support for children learning with technology, this often comes at the expense of the benefits that peer collaboration bri ngs to the learning process, especially in regions where peer-peer collaboration may not be encouraged by teachers such as some cultures in Tanzania.

To begin this thesis, I conducted two research studies to understand how peers support each other in rural, low-resource contexts in Tanzania. In the first study, I deployed a tablet-based educational technology in different social contexts; in school and at home, in the presence or absence of adults, with shared or individual tablets, and in the presence or absence of other knowledgeable children. Based on insights from video observations and interviews, I found that students needed three types of support to successfully engage with the tablet-based learning technology: digital literacy support, application specific support, and domain knowledge support. Peers provided digital literacy, and application support primarily by modeling correct behaviors or selecting answers for their peers. In the presence of a teacher however, peers did not collaborate at all, and depended entirely on teachers for support.

Following these results, I conducted an experimental study where I assigned and trained group leaders to provide adequate peer support while learning with technology. I varied the experimental conditions by making the presence of the leader public in some groups but not in others. I found that group leaders provided adequate and persistent support only in the public condition due to the social expectation of help-giving from their peers. Results from this experimental study showed that with adequate knowledge and training, peers can provide support for each other in this cultural context. However, this arrangement did not promote a culture of help-seeking, help-giving, and collaboration between all members of the group as a whole, and the leader’s new authority position caused them to exhibit behaviors similar to a teacher such as verbally and physically reprimanding students for disturbing the group, interrupting student sessions, and seizing student tablets as disciplinary measures.

To complete my thesis, I explored the design of a system intervention that fostered more equitable helping and collaborative student behaviors by designing a rule-based struggle detection system that automatically detects the kind of support that a student needs, and offers them suggestions to seek for help from another student in the group who has mastered that task. This intervention is based on an Intelligent Novice Model, where students are allowed to engage in productive struggle, and feedback is delayed until the struggle is detrimental to the student experience (unproductive floundering). In this system, every student became a potential helper by navigating an application successfully or scoring well on an activity. I conducted an experimental study in a school located in a rural village in Tanzania over a one-month period, using mixed methods approaches to understand differences in students' helping behavior, their interaction patterns and performance differences on the learning device, how they handle struggle while learning, as well as any changes in their classroom behavior as a result of the increased collaboration between group members during the experimental sessions. Such a study would be incomplete without gaining the perspectives of teachers who have to deploy and incorporate these technologies into their teaching practice. Therefore, I also present insights from observing classrooms and interviewing teachers and administrators in rural Tanzanian classrooms on the socio-cultural ideologies that influence their disposition towards peer-peer collaboration in the classroom. This thesis contributes to research on applying learning science principles in understudied contexts, designing feedback mechanisms for learning systems, and understanding the effects of teaching students such behaviors on normal student-student, and student-teacher interactions.

[1]

144 pages

Thesis Committee:
Amy Ogan (Co-Chair)
Kenneth Koedinger (Co-Chair)
Jessica Hammer (HCII/ETC)
Nicola Dell (Cornell University)

Jodi Forlizzi, 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