How NYS Assessment Constructed Responses Are Graded — And How AI Can Improve the Process

When it comes to measuring student thinking, constructed response questions (open-ended writing, problem-solving tasks, explanations, etc.) offer the richest insights. They showcase not just what a student knows, but how they think. Yet, they’re also the most time-consuming and complex to grade accurately and fairly.

That’s why New York State’s approach to grading stands out as a gold standard: multiple certified raters, a system of checks and balances, and a clear plan for resolving discrepancies. But could schools replicate that model more broadly—without tripling staff time or budgets?

Yes. And AI is the key.

The NYS Multi-Rater Grading Model: A Blueprint for Accuracy

New York’s Grades 3–8 state assessments follow a stringent model:

• Two certified teachers score each open-ended response independently.
• If their scores don’t align, a third certified scorer steps in to determine the final rating.
• Teachers are not allowed to score their own students’ work, minimizing bias.
• Scoring can occur at local, district, regional, or vendor-based centers—but always under strict protocols and training.

This approach drastically reduces subjectivity and ensures that no single person’s bias or fatigue can skew a student’s score.

What Other States Are Doing — And Where Gaps Exist

Let’s compare:

Texas

• Heavy reliance on third-party vendor scorers, many of whom are not active educators.
• Emphasis on speed over depth, often with financial incentives for high-volume scoring.
• Responses evaluated via centralized scoring centers, limiting local educator involvement.

Florida

• Growing reliance on automated scoring engines, especially for writing and performance tasks.
• Inconsistent use of teacher reviewers across schools.
• Concerns about whether algorithms can truly assess nuance, especially for diverse learners.

California

• More localized scoring, especially for performance tasks in CAASPP and Smarter Balanced.
• While this promotes educator ownership, scoring protocols vary and often lack rigorous multi-rater safeguards.
• Inter-rater reliability is a challenge due to inconsistent training and implementation.

The bottom line? New York’s multi-rater model is the most thorough, but also the hardest to scale without additional resources.

Enter AI: The Next Evolution in Fair, Reliable Scoring

What if every school could implement a multi-rater model—without hiring more staff?

That’s what GRADED+, our AI-powered scoring and feedback assistant at Solved Consulting, makes possible.

GRADED+ acts as a second or third set of eyes, giving teachers and school leaders a scalable, reliable way to elevate assessment practices. But beyond just mimicking teacher scoring, AI can analyze student work in deeper, more sophisticated ways than we often expect.

AI Isn't Just Fast—It’s Brilliant

One of the most important things to understand about modern AI tools (like ChatGPT, which powers the core of GRADED+) is this: it’s not just a glorified spell checker. This technology has demonstrated true academic sophistication:

• Passed the LSAT — outperforming most human test takers.
• Scored above average on the SAT, GRE, and GMAT.
• Passed bar exams, medical licensing exams (USMLE), and advanced professional certifications.
• Can accurately solve graduate-level math, science, and economics problems.
• Can evaluate essays using complex rubrics—including tone, argument strength, evidence, and grammar.

In short: AI understands the content. Not just grammar and keywords, but conceptual reasoning, evidence strength, and logic flow.

That means AI can assess student work at a deeper level than most people realize—with consistency and without fatigue.

How GRADED+ Mirrors and Enhances the NYS Grading Model

Whether used as a validation tool for state assessments or a day-to-day assistant for classroom grading, GRADED+ provides:

• Unbiased, standards-aligned scoring
• Instant formative feedback for students
• Teacher calibration support (compare teacher score vs. AI score for professional learning)
• Faster turnaround times for high-quality feedback

Why It Matters at Every Level: From States to Classrooms

State agencies can use AI scoring to support human raters and ensure consistency across large-scale assessments. But the real opportunity lies in empowering classroom teachers—those who grade hundreds of essays, responses, or lab reports each month without a second opinion.

With GRADED+, a teacher in a single classroom can replicate the rigorous scoring checks of a state assessment center. It's like having a certified teaching assistant who never gets tired, never rushes, and knows the content inside out.

And for districts? It’s a way to ensure fairness and equity in grading—no matter which teacher is in the room.

Let’s Rethink Grading Together

We believe every student deserves feedback that’s fair, fast, and insightful. And every teacher deserves tools that support—not replace—their expertise.

With GRADED+, we’re bridging the gap between ambitious assessment practices and real-world classroom demands. Whether you're a district leader, instructional coach, or teacher looking to lighten your load, GRADED+ gives you the power of a third scorer—on demand.

👉 Visit www.solvedconsulting.com/graded to schedule a demo or learn how your school can implement AI-enhanced scoring today.

‍