When several teachers assign the same essay on a shared book, students naturally compare grades. If one teacher is notably more generous than another, the unfairness is obvious and parents notice quickly. A common text such as J. Richard Gott III's "Time Travel in Einstein's Universe" gives a department an ideal chance to align its grading practices.

Consistency begins with a shared assignment sheet and a single rubric that every teacher uses without modification. Even small changes, such as one teacher adding a bonus for creativity, introduce differences that compound across sections. Agree on the rubric as a team and then treat it as fixed for the duration of the unit.

The next step is to calibrate by scoring a small set of anonymous sample essays independently and then comparing results. Differences in scores reveal where the rubric language is ambiguous or where teachers have different assumptions. Discussing those differences openly is the most valuable part of the process.

Collect anchor papers for every score level

Anchor papers are sample essays that exemplify each level of the rubric and serve as reference points during grading. When a borderline essay appears, a teacher can compare it against the anchors instead of relying on memory or mood. Over time, a department can build a library of anchors that new teachers can use to learn the standard.

• Select one anchor essay for each performance level on every rubric row.
• Annotate each anchor to explain why it earned its score.
• Review anchors together at the start of the grading period.
• Return to the anchors whenever a borderline essay appears.
• Update the anchor library each year with new examples.

Use double scoring on a sample

Double scoring a small percentage of essays, with a second teacher grading independently, gives a department objective evidence of how consistent it really is. If scores differ by more than one level on a regular basis, the team knows that more calibration is needed. The sample need not be large, since even ten essays can reveal systematic differences.

Approach the results as information rather than criticism. Some differences reflect genuine ambiguity in the rubric, and fixing the language benefits everyone. Teachers are more willing to participate when the goal is improvement rather than evaluation of individuals.

Watch for drift within a single teacher's grading

Consistency problems are not only between teachers; they also happen within one teacher's stack. Fatigue, mood, and the quality of recently read essays all influence scores, a phenomenon often called drift. A strong essay that follows five weak ones may be graded more generously than it deserves, and the reverse is also true.

Strategies for reducing drift include shuffling the order of essays, grading one rubric row at a time, and taking breaks. Re-scoring a few early essays at the end of the session can reveal how much your standards shifted. These simple practices improve fairness considerably.

Bring in consistent first-pass scoring

A tool that applies the same rubric to every essay in the same way offers a stable reference for human graders. It does not tire, and it does not grade differently on Friday than on Monday. Teachers can compare their own scores with the tool's output to spot where drift may have occurred.

The final decisions should always belong to teachers, who understand context that no tool can capture. Used this way, AI-assisted scoring becomes a calibration aid rather than a replacement. Departments that adopt it thoughtfully often report greater confidence in the fairness of their grades.