|Title:||Developing an Online Learning Environment for Learning Algebra by Teaching a Synthetic Peer|
|Principal Investigator:||Matsuda, Noboru||Awardee:||Texas A&M University|
|Program:||Education Technology [Program Details]|
|Award Period:||4 years (09/01/2018-08/31/2021)||Award Amount:||$1,399,947|
|Type:||Development and Innovation||Award Number:||R305A180319|
Co-Principal Investigator: Huang, Ruihong
Purpose: In this project, researchers will develop and evaluate an online game-like environment for middle school students to solve and learn algebra linear equations by teaching a simulated peer student. Learning by teaching is a promising style of instruction, with evidence supporting that when students engage in peer tutoring there is a benefit for both the tutee and tutor.
Project Activities: In Year 1 and 2, research and development will occur in a research laboratory at Texas A&M University. Following a laboratory usability evaluation study, in Year 3 a series of pilot studies will evaluate the feasibility, fidelity, and promise of proposed environment to improve student learning of algebraic equations.
Products: Researchers will develop a module of Artificial Peer Learning Environment Using SimStudent (APLUS), agame-like environment for middle school students to practice and learn mathematics by collaboratively solving algebraic equations with a simulated peer student.
Setting: The research will take place in a research laboratory at Texas A&M University and in classrooms in suburban schools in medium to-large sized cities in Texas and Pennsylvania.
Sample: For the lab evaluation studies, researchers will recruit participants through Texas A&M University's subject pool. The classroom research in schools will include 120 grade 7 and 8 students and eight algebra teachers.
Intervention: In prior IES and NSF projects, the researchers developed and tested the Artificial Peer Learning Environment Using SimStudent (APLUS), an online environment for students to tutor a simulated peer. In the interaction, students and the simulated peer student solve questions step by step and are provided feedback on the accuracy of the steps performed. In this project, the researchers will develop a module of APLUS within a specific domain, algebraic linear equations for grades 7 to 9. The team will develop a machine learning engine to enhance the student and simulated peers interaction in working through algebraic manipulations.
Research Design and Methods: To develop the environment, the researchers will employ a data-driven system-engineering procedure where iterations of research and development occur in lab evaluation studies and classroom pilot studies until feasibility, usability, and learning aims are met. After development is complete, the researchers will conduct a series of between subjects randomized control trials to examine whether the environment shows promise for improving student learning of algebraic equations. In the studies that will occur over the course of a school year, classrooms will be randomly assigned to implement the tutoring environment or not. Learning outcome measures will be collected from pre-, post-, and delayed-tests to assess the competency in solving equations and conceptual understanding.
Control Condition: The researchers will compare several versions of APLUS to a control group that uses a different version of APLUS to isolate the effect of different approaches and to understand when and how learning by teaching becomes most effective. For each study, students will be randomly assigned into one of the study conditions at the beginning of the school year, with treatment students using the same type of intervention throughout the year.
Key Measures: To iteratively develop and refine the technology, the researchers will use a think-aloud protocol where participants provide feedback while using the prototype. For the pilot studies, the researchers will gather data using a combination of researcher-developed assessments of content and procedural knowledge, will conduct classroom observations, and will use learning processdata to measure the interactions between students and the simulated peer.
Data Analytic Strategies: The research team will employ analysis of variance with between-subjects comparisons to understand the effect of the treatment. Additionally, the researchers will use within-subjects comparisons for all students to examine learning gains by comparing performance at pre-, post-, and delayed-exposure to the intervention and performance over a longer period of time. In addition, researchers will employ correlation and causal analyses to identify key cognitive and background variables that are significantly related to learning outcome measures.
Related Project: Learning by Teaching Synthetic Student: Using SimStudent to Study the Effect of Tutor Learning (R305A090519)