Value-added is unique from other growth measures — it is the only one that isolates the contribution of a teacher or school from other factors that affect student growth.
While value-added models comes in many varieties, all such models use scores from at least two assessments — typically taken a semester or a year apart — from the same group of students. Models differ in the additional types of data they include and the sort of adjustments made. These include:
- Student-level demographic data
- Classroom-level demographic data
- School-level demographic data
- Level of detail in student-teacher links
- Adjustments for measurement error in the tests
- Adjustments for small sample sizes that randomly affect outcomes
The inclusion, or exclusion, of certain variables can impact the results generated by a value-added model and affect the policy decisions based on these results. VARC’s co-build approach encourages careful consideration of what data is included in each model and the potential effects on teachers and schools. This co-build approach, combined with VARC’s professional development, gives districts, administrators, and teachers the ability to interpret and use these results appropriately.