This presentation is grounded in an interdisciplinary perspective that integrates English language arts, engineering problem-solving, and emerging AI/ML literacy within upper elementary education. We approach machine learning not as a distant, opaque “black box,” but as a set of concepts that can be made meaningful to young learners through tangible, hands-on experiences. As AI increasingly shapes the future workforce, early exposure to these ideas is critical for developing foundational AI literacy and empowering students as informed users and future creators.
Our theoretical framework draws on research on tangible interfaces and dual representations to support conceptual understanding of abstract ideas. Decades of mathematics education research demonstrate that manipulatives—concrete, physical representations—help students reason, communicate, and transition toward abstract thinking. Similarly, dual representation theory highlights the power of combining tangible and symbolic experiences to strengthen knowledge transfer and application.
We apply these principles through the use of SmartMotors, a tangible, low-cost interface that allows students to design, train, and test machine learning models in developmentally appropriate ways. By blending language, engineering, and AI/ML concepts through physical interaction, we aim to spark curiosity, deepen conceptual understanding, and foster inclusive access to emerging technologies for all learners.
This study employed a mixed-methods design to examine the integration of SmartMotors technology into upper elementary classrooms and STEM clubs. Participants included both teachers and students who implemented or engaged with the SmartMotors program. Teachers were recruited based on their involvement in pilot implementations and participation in professional development activities. Data collection for teachers consisted of individual interviews or focus groups, each lasting approximately one hour and conducted either in person or virtually via Zoom or Microsoft Teams. Across three sessions, teachers discussed their experiences integrating SmartMotors into their instructional practices, perceived challenges and benefits, and observations of student engagement. In addition, teachers completed multiple Qualtrics surveys administered at several timepoints. These surveys included demographic items (e.g., gender, ethnicity, teaching experience, grade levels, and number of students served) and measured knowledge, confidence, self-efficacy, attitudes, and perceptions of program effectiveness and student learning.
Student data were collected through classroom recordings, interviews, and surveys. Stationary digital recording devices captured students’ interactions with SmartMotors during design challenges and classroom activities. After each activity, teachers conducted brief semi-structured interviews (30 minutes or less) to elicit students’ reflections on their creations, learning processes, and overall experiences. Students also completed pre- and post-surveys that measured machine learning knowledge, attitudes toward AI/ML, self-efficacy, and satisfaction with the program, along with demographic information.
Qualitative data from teacher and student interviews were transcribed and analyzed using thematic analysis to identify patterns in experiences, instructional practices, and conceptual understanding. Quantitative survey data were analyzed using descriptive statistics and repeated measures comparisons to examine changes over time in knowledge, attitudes, and self-efficacy. This multi-source, mixed-methods approach provided a comprehensive understanding of both educator and student perspectives on AI/ML integration and allowed for triangulation across data types.
We expect the results to support our hypotheses. Teachers are anticipated to demonstrate increases in knowledge, confidence, self-efficacy, and positive attitudes toward integrating AI/ML concepts into their instructional practice. We expect qualitative interviews to reveal that teachers view SmartMotors as an accessible, developmentally appropriate tool that fosters student engagement, curiosity, and cross-disciplinary connections. Tangible interfaces are expected to be described as particularly effective for introducing complex machine learning concepts in ways that complement existing STEM and language instruction.
Student data are expected to show measurable gains in machine learning knowledge, self-efficacy, and attitudes. We anticipate significant improvements in students’ understanding of key ML concepts and their confidence in engaging with AI/ML tasks. Quantitative analyses will include dependent-means t-tests and within-subjects ANOVAs to examine changes across timepoints for both teacher and student survey measures. Interview and observation data are expected to highlight students’ enthusiasm, creative problem-solving, and ability to explain their reasoning while interacting with SmartMotors.
For conference audiences, this research offers both practical and theoretical value. It contributes empirical evidence using mixed-methods and quantitative analyses (e.g., dependent-means t-tests and within-subjects ANOVAs) to document measurable gains in knowledge, self-efficacy, and attitudes for both teachers and students. It also provides implementation insights, illustrating how classroom educators can foster innovation, problem-solving, and digital literacy through accessible, low-cost tools.
These findings will be valuable for educators, researchers, curriculum designers, and policymakers seeking evidence-based strategies to broaden participation in AI/ML education, support teacher confidence, and prepare students for future technological landscapes.
National Council of Teachers and Mathematics. Principles to action: Ensuring mathematical success for all. National Council of Teachers of Mathematics., Reston, VA 2014.
2 M. N. Suydam. Manipulative materials and achievement. The Arithmetic Teacher, 33:10-32, 1986.
3 D.H. Clements and M.T. Battisstia. Constructivist learning and teaching. The Arithmetic Teacher, 38:34-35, 1990.
4 J.W. Heddens. Bridging the gap between the concrete and the abstract. The Arithmetic Teacher, 33:14-17, 1986.
5 David H. Uttal, Katherine O’Doherty, Rebecca Newland, Linda Liu Hand, and Judy DeLoache. Dual representation and the linking of concrete and symbolic representations. Child Development Perspectives, 3(3): 156 – 159, 2009. URL: https://srcd.onlinelibrary.wiley.com/
Posters in this theme:
Back
Trips and Tours