Sown to Grow - Enabling Growth Through a Comprehensive Social Emotional Learning and Refection Platform Powered by Machine Learning
Award last edited on: 3/23/2023

Sponsored Program
Awarding Agency
Total Award Amount
Award Phase
Solicitation Topic Code
Principal Investigator
Rupa C Gupta

Company Information

Sown To Grow Inc

515 Crofton Avenue
Oakland, CA 94610
   (415) 745-9465
Location: Single
Congr. District: 13
County: Alameda

Phase I

Contract Number: 91990021C0021
Start Date: 5/3/2021    Completed: 12/31/2021
Phase I year
Phase I Amount
In previous research and development, the company created Sown To Grow, a product for grade nine students to complete a weekly survey and write short reflections describing their social emotional learning, with results provided to educators. In this Phase I project, the research team will develop a new prototype with natural language processing and machine-learning technologies to automatically categorize open-ended student responses, with the insights to be provided to educators to inform intervention if needed. At the end of Phase I, a pilot study will include six educators and their grade 9 students. The researchers will examine whether the prototype generates automated insights that align to the written reflections that students provide, if educators understand the results produced by the prototype dashboard, and whether these results would be useful at providing information that educators could use to support the SEL development of students.

Phase II

Contract Number: 91990022C0041
Start Date: 5/15/2022    Completed: 5/14/2024
Phase II year
Phase II Amount
Purpose: In this project, the team will develop a machine learning algorithm to strengthen an existing social and emotional learning (SEL) intervention for middle school students. Research indicates that adolescents' social and emotional skills strongly correlate with long term academic achievement and life outcomes. However, as students move from elementary to middle to high school, they sometimes experience declines in key social-emotional competencies, and significant gaps exist based on race and socio-economic status. Schools struggle to implement comprehensive SEL programs, and educators are not always trained or equipped to teach these skills. Project Activities:In the Phase I project in 2021, the team developed a new prototype that uses machine-learning functionality to automatically categorize themes in students' open ended emotional reflections on their social and emotional well-being. The team also developed a prototype dashboard that presents educators' insights on the themes to help inform the intervention and the responses provided to students. At the end of Phase I, researchers completed a pilot study with 11 educators and 119 middle school students that determined that the prototype functioned as intended. The machine learning engine successfully categorized student reflections to educators, 85% of educators stated that the information helped them provide more individual feedback to students, and the technology helped deepen relationships between students and teachers. In the Phase II project, the team will fully develop the product components, including optimizing the machine learning algorithm as more data is added to the model, upgrading the user-interface and user-experience for the students and educators, and enhancing the social and emotional learning content and curriculum. The project team will conduct iterative refinements via gathering feedback from educators and students at major production milestones until the product is fully functional. After initial development is complete, a short feasibility study will be conducted at two middle schools to gather feedback for a final development cycle. After development is complete, researchers will carry out an underpowered efficacy study to determine whether the product shows promise in improving student outcomes in social emotional learning. Researchers will implement this study across 10 middle schools, with 60 classrooms participating. Half of the classrooms will be randomly assigned the product and half will not use it. A researcher validated measure of social and emotional learning that is part of the CORE-PACE partnership will be employed to o examine differences among the students treatment and control classrooms from pre- to post-test. Qualitative data will be captured at multiple points to assess educator and student perceptions. Researchers will gather cost information using the "CostOutTool" to determine the cost per student. Product: In previous research and development, the company created Sown To Grow, a product for students to complete weekly surveys and write short reflections describing their social emotional learning. In this Phase II project, the team will fully develop a component of the intervention for middle school students that uses machine-learning to categorize students' survey responses and open-ended written responses. The new component will include an educator dashboard that presents insights and themes on student answers and reflections, and provides educators suggested replies to students based on the information. The project team will also develop educator delivered mini-lessons and professional development for educators.