When a history department assigns the same Pliny essay across several sections, students reasonably expect the grades to reflect the same standard. In practice, scores can vary because teachers interpret criteria differently, weight evidence and style differently, or hold different views of what a strong paper looks like at this level. Calibration is the process of aligning those interpretations. It is time well spent, because it improves fairness and reduces conflict.

Begin with a shared rubric, even if each teacher later adapts it slightly for their classroom. The rubric should describe what proficiency looks like in plain, observable terms. Avoid vague words such as good or adequate, and prefer descriptions that point to features of the writing, like "explains the relevance of at least two quotations to the thesis."

Collect a small set of anonymous student papers on the assignment, ideally covering a range of quality. Five to eight papers are usually enough for a productive session. Remove names and any identifying details, and distribute copies in advance so teachers can read and score them on their own before the meeting begins.

Run the Calibration Meeting

At the meeting, compare scores paper by paper, starting with those where disagreement was largest. Ask each teacher to explain which evidence in the paper led to the score, referring to the rubric language. These discussions often reveal that two teachers read the same descriptor in different ways, and the group can then agree on a shared interpretation and record it.

• Score sample papers independently before discussing
• Compare results and flag the rubric rows with the greatest disagreement
• Discuss the evidence in each paper that justified each score
• Write down agreed interpretations of ambiguous descriptors
• Select anchor papers to represent each score level for future use

Create Anchor Papers and Commentary

Anchor papers are annotated examples that show what each score level looks like in practice. With brief commentary explaining why each paper earned its score, they become powerful reference tools for new teachers, substitutes, and teaching assistants. They also help students, who can see what high-quality work looks like and compare it with their own.

Update the anchors periodically as the assignment evolves or as the department's expectations shift. A small library of anchor papers for each major writing task, including the Pliny essay, becomes a valuable shared resource. Storing them in a central location ensures that everyone can access them and that the effort invested in calibration continues to pay off year after year.

Check Consistency During Grading

Calibration does not end when the meeting does. Midway through the grading period, exchange a handful of graded papers among teachers and compare scores again. This quick spot-check can catch drift and give teachers a chance to adjust before the grades are finalized and returned to students.

Some departments also track statistics, such as the average and distribution of scores in each section. Large differences are not proof of unfairness, since classes vary, but they can prompt useful questions about instruction and grading. The goal is not to force uniform outcomes but to ensure that the standard being applied is consistent and defensible.

Support Calibration With Technology

AI grading tools can serve as an additional reference point during calibration. Running the same sample papers through a tool configured with the shared rubric can show where its scoring aligns with the group and where it diverges, which often surfaces unclear rubric language. This is useful as a diagnostic, even for departments that never plan to use the tool for live grading.

If a department does adopt AI-assisted grading, calibration becomes even more important. Teachers should review the tool's output against anchor papers, document the settings used, and revisit them regularly. Treating the technology as another grader to be calibrated keeps human judgment at the center and ensures that standards remain consistent across the department.