When a literature course has multiple sections, students quietly compare notes. One section might hear that a particular instructor gives easier grades on essays about Guo Moruo's The Goddesses, while another is known for harsh scoring on close reading. These perceptions, whether accurate or not, undermine confidence in the course and create pressure on instructors to justify their marks.

Grade drift happens for ordinary reasons. Different instructors emphasize different skills, interpret rubric language in their own way, and bring different tolerances for errors to their reading. Without a deliberate process to align expectations, even a well-written rubric will be applied unevenly.

The most reliable remedy is calibration, a structured process in which instructors grade the same essays and discuss their scores. Select three or four anonymous papers that span the range of quality, have each instructor score them independently, and then compare results. The conversation that follows often exposes differences in how the rubric is understood and gives the group a chance to resolve them.

Choose anchor papers carefully

Anchor papers serve as reference points for each score level, so they should be chosen with care. A good set includes a clear top-level essay, a solid middle essay, and a weaker essay that shows typical problems such as summary instead of analysis. Annotating each paper with notes about why it earned its score gives instructors a model to return to during grading.

• Select anchor papers that reflect common student work, not extreme cases.
• Annotate each anchor with the rubric rows it satisfies or misses.
• Share anchors with all instructors before grading begins.
• Revisit and update anchors each term as assignments change.
• Consider sharing anonymized anchors with students to clarify expectations.

Check agreement during the grading period

Calibration should not be a one-time event. During the grading period, instructors can exchange a small sample of graded papers and spot-check each other's scores. A quick review of five or six papers per instructor can reveal whether someone has drifted from the shared standard and allows for adjustment before grades are released.

Tracking score distributions across sections is another useful check. If one section's average is much higher than another's without a clear reason, it is worth looking at why. The goal is not to force identical averages but to confirm that differences reflect real differences in student performance rather than differences in grading.

Standardize feedback as well as scores

Consistency is about more than numbers; students also notice differences in the quality and tone of feedback. One instructor may provide detailed marginal comments while another offers only a brief summary. Agreeing on a minimum standard, such as a comment on thesis, a comment on evidence, and a next step, ensures that every student receives useful guidance.

Shared comment banks and templates can support this standardization. They give instructors a starting point for common issues and save time without sacrificing personalization. Departments that invest in these resources often find that new instructors come up to speed much more quickly.

Use technology to support alignment

AI-assisted grading tools can help enforce consistency by applying the same rubric language to every essay in every section. They can also highlight cases where a score and its comments seem misaligned, prompting a second look. These features do not replace instructor judgment, but they provide an additional layer of checking that is hard to achieve manually.

The most important factor is still the human agreement behind the rubric. Technology can apply a standard consistently, but it cannot decide what the standard should be. Departments that invest in shared understanding first will find that any tool they adopt works much better.