“How to grade a paper” covers two different tasks that most guides blur together: scoring parts with a right answer, and evaluating parts that need judgment. A paper with short answer questions and an essay section needs both, and grading them the same way is how inconsistent scores happen.
This walks through both, from the first objective answer to the final letter grade, plus keeping feedback useful enough that students revise from it. For essay feedback specifically, an AI rubric generator is the tool most of this depends on.
Objective answers are close to a solved problem, you check against a key and count wrong answers against the number of problems. Subjective work has no key, which is why the same paper can look like a B to one grader and a B-plus to another.
The gap isn’t effort, it’s standard. Two graders who haven’t agreed on what “strong evidence” looks like will drift apart on the same stack, which is a rubric problem, not a speed problem, different from how to grade essays faster.
This matters most on papers mixing both question types. A teacher who scores the essay section first, while still calibrating their standard, applies a looser or stricter read to what follows, without meaning to. That drift shows up in AI vs human grading accuracy.
1. Decide what kind of paper you are grading
Separate the objective sections from the subjective ones before you start, a paper mixing short answer questions with a writing assignment needs two approaches, not one pass. Objective answers get checked against a key; subjective sections need a rubric first, or standards drift partway through the stack.
Decide the grading order for the whole stack too, not just within one paper. Grading every student’s answer to question one, then moving to question two, keeps your standard steadier than grading paper by paper straight through.
2. Score the objective parts first
Multiple choice, fill-in-the-blank, and short answer questions with one correct answer go fastest and clear the most straightforward part of the stack. Count wrong answers against the number of problems and record the raw score before anything subjective. If you want software for this part, see AI grading software.
3. Apply a rubric to student essays
For student essays and other open-ended writing, score against a rubric tied to the assignment’s actual learning objectives, not a general impression. Read for the criterion you’re scoring, not the paper as a whole, since that’s how two graders land on different scores.
If you don’t have a rubric for this assignment yet, build one before you score the first paper, not partway through. A rubric written mid-batch tends to reward whatever you happened to notice first. Checking quality grading accuracy periodically catches this drift.
4. Leave feedback that supports revision
Comment on what a student can act on, not everything you noticed. One clear note tied to the rubric is more useful than ten small corrections scattered across the margins, especially if the student will revise before a final grade is recorded.
A good comment names the specific fix: “your evidence in paragraph two doesn’t match the claim in paragraph one.” A vague note like “needs more detail” leaves a student guessing at what to actually change.
5. Calculate the final grade
Combine the objective score and the rubric-based score using the assignment’s weighting, then convert to a letter grade on the same scale you’d use for any assignment. Write down the math, not just the result, so a parent asking how you got there gets a real answer.
Decide your policy on partial credit and extra credit before you start, not case by case. A consistent policy applied to every paper is what actually holds up if a grade gets questioned later.
6. Track what is still ungraded
Keep a running count of what’s ungraded so a stack doesn’t quietly become a two-week backlog. If you’re grading inside Canvas SpeedGrader, use its own progress view instead of a separate tracker, and flag anything uncertain for office hours. A growing backlog is covered in teacher time management.
EnlightenAI helps teachers deliver instant, rubric-aligned AI writing feedback so students can practice, revise, and improve faster. It's a simple way to start grading essays more efficiently.
Every step above is manageable for one paper. The real test is whether the tenth paper gets the same standard as the first, and whether a colleague grading the same rubric agrees. That’s what district grading at scale actually requires.
If keeping this process consistent across a full stack is the hardest part to sustain, that’s the specific gap an AI grading tool is built to close.
A second pair of eyes helps most here. Ask a colleague to grade three of your papers blind and compare scores before trusting your own calibration on a full stack.
