The research behind CoGrader

CoGrader expands teachers’ capacity to give high-quality, rubric-aligned feedback across the writing process. Students write, get criterion-specific feedback, revise, and write again. Teacher judgment stays in place at every step.

Published September 4, 2026. Last updated September 8, 2026.

U.S. Department of Education IES research award Award R305J250071

How CoGrader is built

Three commitments shape the product. Each one is a product rule.

Teacher-defined criteria
The teacher picks or writes the rubric. CoGrader scores against that rubric, not a generic standard.
AI-supported evaluation
CoGrader drafts a score and criterion-level feedback for each submission.
Teacher-controlled judgment
The teacher reviews, edits, and approves before anything reaches a student.

This matches the research consensus. Automated writing evaluation works best next to teacher-led instruction, not in place of teacher judgment.

What the evidence supports

Every claim we make, how far the evidence goes, and where it comes from.

Claim	Status	Basis
Frequent writing and revision improve student writing	Supported by decades of research	Graham & Perin (2007); IES practice guides (2012, 2016)
Timely, criterion-specific feedback improves writing	Supported	Graham, Hebert & Harris (2015); Hattie & Timperley (2007)
Automated writing feedback can help when teachers integrate it into instruction	Supported *	Tseng et al. (2026) meta-synthesis; Fleckenstein et al. (2023)
Teacher oversight and rubric alignment are recommended	Supported	Tseng et al. (2026); Zheldibayeva et al. (2026)
CoGrader’s STAAR scores agree closely with TEA raters	Benchmark available	CoGrader STAAR validation (June 2026)

* The effect depends on how the teacher uses the feedback. The studies support automated feedback as a supplement to teacher-led instruction, with the teacher setting the criteria and reviewing what students see.

Full literature review: The research behind AI grading

Studies

Two CoGrader studies and one independent, peer-reviewed study. Each one names its method, its date, and what it does not show.

Technical report June 2026

STAAR constructed response scoring validation

CoGrader’s scores landed within one point of TEA raters in 98.5% of 600 responses across eight scoring modules.

Benchmarked against the scoring guides TEA publishes to train its own raters. Not peer reviewed.

600
TEA-scored responses: 8
scoring modules: 98.5%
within one point

Read the whitepaper

Field study Spring 2026

CoGrader in the classroom: a Texas ISD after one year

Site visit, classroom observation, and group interviews with teachers and district leaders. Five themes.

Describes practitioner experience as educators reported it. It does not measure effects.

2
campuses visited: 4
educator groups: 5
themes

Read the brief

Peer-reviewed study 2024

EFL teachers’ perceptions of an AI grading tool (CoGrader) in English writing assessment at Saudi universities

Independent of CoGrader. Examined how university EFL teachers experienced the tool in writing assessment, through an activity theory lens.

Alsalem, M. S. (2024). Cogent Education, 11(1), Article 2430865.

Read the paper

IES award R305J250071

Funded by the U.S. Department of Education’s Institute of Education Sciences

CoGrader’s research and development is funded by the U.S. Department of Education’s Institute of Education Sciences (IES) through award R305J250071. The award is part of From Seedlings to Scale, a program in IES’s Accelerate, Transform, Scale initiative, in the focus area Seamless Personalized Education Experiences Delivered at Scale.

Program: From Seedlings to Scale (S2S), Accelerate, Transform, Scale (ATS)
Focus area: Seamless Personalized Education Experiences Delivered at Scale (SPEED at Scale)
Awardee: Modern Learner Media, LLC (CoGrader)
Principal investigator: Gil Quadros Flores
Award period: September 30, 2025 to September 29, 2026

View the award record on ies.ed.gov

What the award has funded so far

Ran focus groups with teachers, school leaders, and students on grading burden, feedback, and where AI should and should not act.
Made a multi-day research site visit to a Texas ISD to observe classrooms and interview teachers and district leaders.
Built and tested rubric-grounded evaluation, drafted feedback, teacher verify-and-correct, and delivery through the LMS teachers already use.
Wrote CoGrader’s research and evidence framework and a research agenda that follows IES’s Standards for Excellence in Education Research (SEER).

IES funding is not an endorsement of CoGrader by IES or the U.S. Department of Education. It funds the research. We will publish the findings either way.

The research reported here was supported by the Institute of Education Sciences, U.S. Department of Education, through Grant R305J250071 to Modern Learner Media, LLC. The opinions expressed are those of the authors and do not represent views of the Institute or the U.S. Department of Education.

CoGrader theory of change

AI-powered feedback that saves teachers time and improves student writing outcomes.

The problem

Teachers run out of time for feedback

Teachers lack sufficient time to provide timely, high-quality, personalized feedback on student writing, resulting in delayed feedback cycles, reduced writing practice, and lower student writing proficiency, particularly for historically underserved learners.

Under these conditions

What has to be in place

School infrastructure and LMS compatibility
Teacher completion of CoGrader Certification
Student access to devices and digital tools
Policy environment around AI use in education
Ongoing leadership support and communication
Equity in access across student populations

CoGrader provides

AI-assisted grading that reduces teacher grading time
High-quality, rubric-aligned, customizable feedback suggestions
Seamless integration into teacher workflows (LMS, existing routines)
Real-time analytics on student performance and writing patterns
Student-facing tools for iterative feedback and revision
OCR scanning of written drafts in home language to allow feedback and suggestions for improvement starting at draft one

Then

Through these mechanisms

Teacher-level mechanisms

Increased available instructional time
Increased frequency and quality of feedback
Greater ability to differentiate instruction
Improved data-informed instructional decisions
Teachers get time that they apply to planning, personalization, and other high-impact work

Student-level mechanisms

Reduced feedback latency (from weeks to days)
Increased opportunities for revision and practice
Greater clarity on expectations and improvement pathways
Increased engagement and ownership of writing
Increase in writing practice via writing assignments due to available grading and feedback support

System-level mechanisms

More consistent and objective grading practices
Improved visibility into student learning trends
Enhanced capacity for targeted intervention
Scalable support for multilingual learners through OCR and AI feedback

Leading to

These outcomes

Short-term outcomes 0 to 12 months

Increased frequency of writing assignments
Reduced turnaround time for feedback
Increased teacher use of formative assessment data
Improved student engagement in writing tasks
Increased teacher satisfaction and reduced grading burden
Teachers get time that they apply to planning, personalization, and other high-impact work

Intermediate outcomes 1 to 3 years

Improved quality of student writing (rubric-aligned measures)
Improved instructional alignment to student needs
Increased student revision behaviors
Reduced disparities in access to high-quality feedback
Increase in writing practice via writing assignments

Long-term outcomes 3+ years

Increased student writing proficiency
Improved academic achievement across content areas
Improved teacher retention and reduced burnout
Scalable model for personalized learning at the classroom and system level
Increased participation in open-ended accountability assessments

This theory of change articulates how CoGrader is expected to improve writing instruction and outcomes through AI-assisted feedback that empowers teachers and engages students. It states hypotheses. The IES-funded studies are designed to test them.

ESSA evidence status: Tier 4, demonstrates a rationale

We believe CoGrader meets the requirements for ESSA Tier 4. Two things support that view. Tier determinations are made by state and local education agencies, not by vendors.

A research-based logic model
The product is built on the research in the ledger above: frequent writing, timely criterion-specific feedback, rubrics, and revision, with the teacher in control. The full theory of change is published on this page.
See the theory of change
An active IES-funded study
Award R305J250071 funds the research and development, including the studies that will test CoGrader’s effect on student writing.

New to the tiers? Read our guide to ESSA tiers of evidence

Frequently Asked Questions

Straight answers about the evidence behind CoGrader.

Is CoGrader research-based?

CoGrader’s design is evidence-based. It is built on decades of research on writing practice, timely feedback, rubrics, and revision. Its research and development is funded by a U.S. Department of Education IES award. No causal study of CoGrader’s effect on student writing exists yet. That study is in progress.

How accurate is CoGrader?

In a June 2026 validation study, CoGrader’s STAAR scores landed within one point of the official TEA score in 98.5% of times. Read the whitepaper.

Does CoGrader meet ESSA evidence tiers?

We believe CoGrader meets the requirements for ESSA Tier 4, “demonstrates a rationale.” It has a research-based logic model, published on this page as the theory of change, and an active IES-funded study. Tier determinations are made by state and local education agencies, not by vendors.

Is CoGrader backed by the Department of Education?

Department of Education IES award R305J250071 funds CoGrader’s research and development under From Seedlings to Scale, an IES program. The award funds the work of defining the problem, refining the product with educators, and planning the studies that will test its impact. It is research funding. It is not an endorsement by IES or the U.S. Department of Education.

Is CoGrader peer reviewed?

The STAAR validation is a CoGrader technical report. It is not peer reviewed. One independent, peer-reviewed study (Alsalem, 2024) examined teachers’ experience with CoGrader in university writing assessment.

Does CoGrader replace the teacher?

No. The teacher sets the criteria, reviews every score and comment, and decides what students see.