The unit wraps up, five sections turn in the same essay on I Know Why the Caged Bird Sings, and you settle in to grade. By the third stack, something has shifted. You are a little more impatient, a little more forgiving, or a little more focused on things you did not notice earlier. Scoring drift is normal, and it is one of the least discussed fairness problems in secondary English.

The causes are ordinary. Fatigue lowers attention. A very strong essay makes the next one look weaker by comparison. A very weak one does the opposite. Even the time of day can shift how generous a reader feels.

Students notice inconsistency more than teachers think. When two friends in different periods compare grades on the same assignment and find a gap, trust erodes fast. A grade is only meaningful if students believe it would be the same regardless of who read it or when.

The good news is that consistency is largely a matter of process, not talent. A few structural habits reduce drift substantially.

Habits That Reduce Scoring Drift

The most effective single change is to score by criterion instead of by essay. Read every thesis first, then every use of evidence, and so on. That way you apply each standard across the whole set with a consistent mindset.

• Choose three anchor essays at different score levels before you begin and refer back to them often.
• Mix up the order of sections, or reverse the alphabet on the second pass, to spread fatigue evenly.
• Re-score five essays from the first batch at the end to check whether your standards moved.
• Take a break every 20 to 30 essays instead of pushing through in one sitting.
• Record borderline decisions in a short log so you can apply the same reasoning later.

Why Written Descriptors Matter

Vague rubric language invites drift because each grading session reinterprets it. Descriptors that spell out observable features, such as the number of pieces of evidence or the presence of an explanation for each, leave less room for shifting standards. Rewrite any descriptor you find yourself interpreting differently on different days.

Sharing the rubric with students before they write helps too. It creates a public standard that you are accountable to, which makes it harder to drift without noticing.

Using a Second Reader

A second reader is the gold standard, but it is rarely practical for 150 essays. A realistic version is to have a colleague score a small sample, perhaps ten essays, and compare results. The conversation about where you disagreed is often more valuable than the scores themselves.

That sample also serves as a calibration check for the department. If two teachers consistently differ on a certain row of the rubric, the descriptor probably needs to be clearer.

Letting Software Act as a Steady Reader

AI grading tools apply the same criteria to the first essay and the last, without fatigue. A tool like GraideMind can produce rubric-based scores and comments for every submission, giving you a consistent baseline to review. You can then focus on where your judgment should override the tool.

Many teachers find that comparing their own scores to a consistent first pass reveals patterns they did not know they had, such as grading harder on organization late in the evening. That awareness alone improves fairness over time.