Literary analysis essays on a beloved novel are hard to grade evenly. Students care about the book, so their writing is often sincere, but sincere is not the same as analytical. A rubric built for this specific kind of essay helps you reward the second without punishing the first.

Start with four or five criteria and resist the urge to add more. Thesis, evidence, analysis, organization, and conventions cover most of what matters in an essay on Perks. More rows make the rubric feel thorough, but they also make it slower to use and harder for students to read.

Weight the criteria to match what you taught. If the unit focused on close reading, analysis should carry more points than conventions. Students notice where the points sit, and they will put their effort there.

Write the descriptors in plain language a sophomore can follow. "Explains why the evidence supports the claim" is clearer than "demonstrates sophisticated inferential reasoning." If a student cannot tell what a level looks like, the rubric is not doing its job.

The Five Criteria That Do the Heavy Lifting

Each criterion should describe something observable on the page. Avoid rows that measure effort or personality, because those are impossible to score consistently. Ask yourself whether two teachers reading the same paragraph would land on the same level.

• Thesis: makes an arguable claim about the novel rather than restating the prompt.
• Evidence: selects specific moments from the text and introduces them clearly.
• Analysis: explains how the evidence supports the claim and what it reveals about Charlie, his relationships, or the form.
• Organization: moves through ideas in a logical order with transitions that do real work.
• Conventions: uses accurate grammar, punctuation, and formatting, including correct handling of quoted passages.

Score Descriptors That Students Can Use

Four levels are usually enough. Fewer than that and you cannot separate strong work from adequate work; more than that and the differences between levels turn into hair splitting. Keep the language parallel across levels so the only thing that changes is the quality.

Pull one or two real student paragraphs for each level and keep them as anchors. When you feel unsure mid-stack, compare the paper in your hand to the anchor rather than to the last essay you read. This one habit fixes a lot of drift.

Calibrating Before You Grade the Whole Stack

Score five essays, set them aside for a day, and score them again. If your numbers moved, the rubric language needs tightening. This takes an hour and saves you from redoing dozens of papers later.

If you work in a department, do this together. Each teacher scores the same three essays, and you compare. The conversation that follows usually shows exactly where the rubric is vague.

Using the Rubric With AI Feedback Tools

A clear rubric also makes AI grading tools far more useful. A platform like GraideMind grades against the criteria you provide, so vague rows produce vague feedback and sharp rows produce feedback students can act on. The quality of the output tracks the quality of the rubric you feed in.

Whatever you use, read a sample of the results yourself and adjust. The rubric is your standard, and the tool is there to apply it at a speed no one can match by hand. Treat the first batch as a test of the rubric as much as the tool.