Value-Added Teacher Evaluation
Also known as: Teacher Value-Added Models, VAM for Teachers, Teacher Effect Estimation, Value-Added Teacher Accountability
Value-added teacher evaluation uses longitudinal student test scores to estimate how much individual teachers contribute to their students' achievement growth, net of what students brought into the classroom. Statistically it applies value-added and mixed-model machinery — controlling for prior achievement and student characteristics, then treating each teacher's residual contribution as an effect to be estimated. Pioneered in Tennessee's TVAAS and scrutinized in a large methodological and policy literature, it became central, and controversial, in teacher accountability.
Key highlights
- Adjusts for student intake, offering a fairer comparison than raw class averages.
- Quantifies how much teachers differ in their contribution to measured achievement growth.
- Uses mixed-model shrinkage to stabilize noisy individual estimates.
- Provides an objective, data-based input to complement classroom observation and other evidence.
Intuition
This section is available to Pro members. Upgrade to Pro
How it works
This section is available to Pro members. Upgrade to Pro
When to use it
Value-added estimation of teacher effects is appropriate as one input among several for low-stakes feedback, professional development, and research into teacher quality and its distribution, where its uncertainty can be respected. It requires longitudinal, vertically comparable test scores reliably linked to teachers and adequate students per teacher. It is widely regarded as too imprecise and assumption-dependent to be the sole or dominant basis for high-stakes individual personnel decisions, a caution stressed by professional statistical associations. It estimates contributions to tested outcomes only, says nothing about untested goals, and assumes students are not sorted to teachers in ways the model fails to capture.
Strengths & limitations
- Adjusts for student intake, offering a fairer comparison than raw class averages.
- Quantifies how much teachers differ in their contribution to measured achievement growth.
- Uses mixed-model shrinkage to stabilize noisy individual estimates.
- Provides an objective, data-based input to complement classroom observation and other evidence.
- Estimates are imprecise and unstable year to year, with wide confidence intervals for individual teachers.
- Nonrandom sorting of students to teachers can bias effects in ways adjustment may not remove.
- Captures only tested subjects and outcomes, ignoring much of what teaching aims to achieve.
- Sensitive to model choice (gain vs. covariate-adjustment vs. layered), test scaling, and persistence assumptions.
Common pitfalls
This section is available to Pro members. Upgrade to Pro
Applications
This section is available to Pro members. Upgrade to Pro
Frequently asked
How does value-added teacher evaluation relate to value-added modeling generally?
It is the application of value-added modeling specifically to individual teachers rather than schools or programs. The same intake-adjusted, often multilevel, machinery is used, but estimating effects at the teacher level is far more demanding because each teacher is observed on few students, sorting is harder to rule out, and the stakes are personal. The general entry covers the modeling framework; this entry focuses on its use, assumptions, and controversies for teacher evaluation. See the related Value-Added Modeling entry.
Why is using value-added for high-stakes teacher decisions controversial?
Because individual teacher estimates are noisy and unstable: the same teacher can rank very differently across years or tests, confidence intervals are wide, and nonrandom student sorting can bias results. The American Statistical Association and others have warned that value-added measures should not be the primary basis for consequential personnel decisions, recommending they be used cautiously alongside other evidence. The statistics can inform judgment, but treating an imprecise estimate as a definitive verdict on a career is the core objection.
What is the 'persistence' or layered-model issue?
A student's score this year reflects not only the current teacher but lingering effects of past teachers. If the model credits all of this year's growth to the current teacher, it misattributes carryover. Sanders's layered model tries to accumulate and separate teacher contributions across years, and other models make explicit assumptions about how prior-teacher effects persist (fully, partially, or not at all). These persistence assumptions materially change teacher estimates and are difficult to verify, which is one reason results are sensitive to model choice.
Sources
- 1.McCaffrey, D. F., Lockwood, J. R., Koretz, D., Louis, T. A., & Hamilton, L. (2004). Models for value-added modeling of teacher effects. Journal of Educational and Behavioral Statistics, 29(1), 67–101.
- 2.Sanders, W. L., & Horn, S. P. (1994). The Tennessee Value-Added Assessment System (TVAAS): Mixed-model methodology in educational assessment. Journal of Personnel Evaluation in Education, 8(3), 299–311.
You have read it. What now?
Cite this page
ScholarGate. (2026, June 22). Value-Added Teacher Evaluation. ScholarGate. https://scholargate.app/education/value-added-teacher-evaluation