A Blueprint for Scaling Tutoring and Mentoring Across Public Schools (cost model)
Kraft, M. A., & Falken, G. T. · 2021
grade Dreviewindependentfailed
Sample
Policy cost model
Population
US K-12 (projection).
Design
Bottom-up cost architecture for scaled tutoring via tiered tutor pipelines.
Key findings
The reference cost architecture: per-pupil cost is driven almost entirely by tutor type and ratio. Cites the NCLB SES failure ($2B/yr vendor tutoring: low take-up, ~0.06 effects) as the cautionary precedent.
Genetic confound
N/A.
Replication notes
Its scale-effect optimism is partly corrected by the authors' own 2024 meta (50-70% attenuation at the scales envisioned).
DOI / URL
10.1177/23328584211042858
Effects
| Outcome | Metric | Value | Measure | Timing | Vs | Horizon | Class |
|---|---|---|---|---|---|---|---|
| Modeled cost per pupil per year | USD | $662 elementary / $927 middle / $1,486 high school (volunteer/stipend pipelines); $2,500-3,800 for paid professional 2:1 models | unknown | projection | none | not-applicable | domain-skill |
| National cost by targeting | USD | $4.8B (bottom-quartile K-8) to $50.1B (universal K-12) | unknown | projection | none | not-applicable | domain-skill |
Cited by
- Tutoring — the honest effect, the Bloom 2-sigma myth, and what survives scalestrong supportconf: highgc: low