Essays on The Great Divorce are harder to grade than they first appear, because Lewis wrote a short dream vision that invites interpretation on nearly every page. Students often drift into plot summary, moral lecturing, or vague claims about heaven and hell without ever building a real argument about the text. A teacher facing forty papers needs a way to separate careful interpretation from confident-sounding guesses, and that takes one standard applied the same way to every paper.

AI grading tools are useful here because they apply a rubric without getting tired at paper thirty-five, when human graders often grow either harsher or more forgiving than they were at paper five. A tool that reads each essay against the same criteria can flag missing evidence, weak thesis statements, and unexplained quotations quickly. That leaves the teacher free to spend attention on the judgment calls that actually need a human reader.

The most common mistake teachers make is treating AI feedback as a finished verdict instead of a first pass. A better approach is to let the tool draft comments against your rubric, then read the essay yourself and adjust anything that misses the assignment's intent. Over a few units, you learn which comments you almost always accept and which ones you rewrite, and that tells you how to tighten your rubric language.

Start With a Rubric Built Around the Text

A generic literary analysis rubric will miss what makes this book distinctive, so tie your criteria to the way Lewis actually works. Students should be judged on whether they explain how a specific ghost, such as the one who clings to a lizard on his shoulder, illustrates a particular spiritual habit rather than just retelling the scene. Criteria that reward interpretation of image and symbol will steer students away from summary far better than a vague line about analysis.

• Thesis makes a debatable claim about choice, pride, or self-deception rather than restating the plot
• Evidence comes from specific scenes and is quoted or paraphrased accurately
• Analysis explains how Lewis's imagery produces meaning instead of only naming a theme
• Organization moves from claim to evidence to reasoning in each body paragraph
• Language stays precise, with correct handling of terms like allegory, dream vision, and ghost

Calibrate the Tool Before You Trust It

Before grading a full set, run three or four essays you have already scored by hand, ideally one strong, one middling, and one weak. Compare the tool's scores and comments against your own and note where it is too generous or too strict. If it rewards length over insight, or misses that a student has misread the ending, adjust your rubric wording or instructions until the results line up with your judgment.

This calibration step also shows you whether your own rubric is clear. When the tool scores an essay differently than you expected, the gap sometimes reveals an ambiguous criterion, such as an undefined phrase like "strong analysis." Fixing that language improves the assessment for students and makes your feedback easier to defend if a parent or a colleague asks how a grade was determined.

Keep the Feedback Specific to the Book

Feedback that could be pasted onto any novel essay teaches students very little. A useful comment on a Great Divorce paper names the moment in the text, explains what the student did well or missed, and suggests a next step, such as connecting the bishop's love of open-ended discussion to the book's warning about comfortable abstraction. Comments like these show students that a reader engaged with their actual argument.

Encourage the tool to phrase comments as questions or prompts when a student's idea is promising but underdeveloped. A note asking why the narrator's guide, George MacDonald, speaks so bluntly about choice pushes a student to think, while a note saying "expand this point" pushes nobody anywhere. Teachers who review and lightly edit these comments end up with feedback that sounds like them and still arrives in days rather than weeks.

Protect Student Growth Across the Unit

The real payoff of faster grading is more feedback cycles, not more free evenings. If a draft comes back in two days, a student can revise their thesis while the reading is still fresh, and the second draft usually shows far more improvement than a rewrite done a month later. Teachers can plan a short draft, a quick round of feedback, and a final essay within a single unit.

Track patterns across the class as well, since a good grading workflow reveals what the whole room is struggling with. If many students treat the ghosts as simple villains instead of people making small daily choices, that is a reteaching opportunity rather than thirty separate mistakes. Using AI to surface those patterns turns a grading pile into useful information about your instruction.