UpGrade is an open source environment to create and manage online randomized controlled trials in educational software. Any EdTech client app can deploy the UpGrade service in its infrastructure.
- Learning Science and Continuous Improvement in EdTech software
- Current and prior studies have used UpGrade with Carnegie Learning's MATHia software, and Playpower's Battleship Numberline game
- We can use A/B testing at scale to scientifically and measurably improve EdTech software
- UpGrade is specific for educational experiments in schools and classroom settings
- UpGrade's support of group random assignment allows us to run experiments in educational settings while maintaining consistent learning experiences within a group
- Reduce teacher burden and possible class disruption by maintaining group consistency
- Allows users to set consistency parameter settings for managing experiment behavior in adaptive educational software, when students reach the content of interest at different times
- How can UpGrade help different teams in your organization?
- Research Teams
- Design and deploy experiments easily
- Monitor experiments using data feedback loops
- Automatically start and end experiments
- Export data for further analysis
- Run multi-site experiments
- Engineering Teams
- Use feature flags to release features over time
- QA new features in production before releasing them to everyone
- Easily keep track of the errors
- A structured way to experiment with features
- Product Teams
- Run A/B tests with consistent user experiences at the individual, class, school, or district level
- Test how new effective new features
- Exclude or manually assign users from the field tests
- See how experiments run in a demo mode
How to start using UpGrade?
Developer: Read our Quick Start guide that will walk you through installing and running UpGrade on your local computer or in the cloud.
Researcher: Read the Experimentation in UpGrade section, then the Example sections to see how UpGrade can help you improve learning experiences at scale.