Character analysis is the most common essay task attached to Stone Cold, and it is also one of the easiest to mark inconsistently. Students tend to describe what Link does and what Shelter did, then call the result analysis. A rubric that does not distinguish between description and interpretation will give similar scores to very different pieces of thinking. Getting the criteria right at the start saves hours of rationalizing marks later.

The central difficulty in this unit is that the two main characters are presented through very different voices. Link speaks in a guarded, wary first person that reveals his fear and stubborn pride, while Shelter's sections expose a man who sees himself as a tidy-minded soldier doing a necessary job. A strong essay notices that both narrators are speaking directly to us and asks what each voice is trying to make the reader feel. Weaker essays treat them as plot devices and move on.

Before building any descriptors, decide what your department actually values in a character essay. Some teachers want a clear thesis about how Swindells builds sympathy for Link, others want attention to how Shelter's self-justification disturbs the reader. Neither choice is wrong, but mixing them in one rubric produces confusion for students and markers alike. Write down the one or two big ideas that every top-band response should engage with.

Four Criteria Worth Scoring Separately

Separating the rubric into distinct strands prevents a polished writer from earning a high mark with thin ideas. Scoring interpretation, textual evidence, analysis of language and structure, and written accuracy independently gives students a clearer map of what to improve. It also makes moderation easier, since colleagues can disagree about one strand without reopening the whole mark. Four strands is usually enough, and more than six tends to overwhelm both students and teachers.

• Interpretation: offers an original, arguable view of Link or Shelter beyond what happens in the plot
• Evidence: selects precise, brief quotations and explains how they support the claim
• Writer's methods: comments on voice, structure, word choice, and the effect on the reader
• Context: connects the characters to wider ideas about homelessness, invisibility, and power
• Accuracy: sustains clear, formal writing with few distracting errors

Writing Level Descriptors That Students Can Use

Descriptors work best when they use observable behavior rather than vague praise. Instead of saying a top response shows insight, say that it explains why Link's refusal to accept help from strangers reveals both his pride and his fear of being owed something. Instead of saying a weaker response lacks detail, say that it names an event without commenting on any word the author used. Students can act on language this concrete without needing a conference.

Give each level a short sample sentence if space allows. A sentence like "Link seems sad" sits at the foundation, while a sentence tracing how his short, clipped statements imply that he is protecting himself from further hurt sits higher. Showing the gap between the two makes the rubric teachable rather than merely evaluative. Many teachers keep a bank of these model sentences and refresh them each year.

Common Marking Errors in Character Essays

One frequent mistake is rewarding length. A long essay that retells the novel in order can look impressive on the page while saying very little about either character. Another is overweighting Shelter, because his chapters are dramatic and easy to quote, which can leave Link's more subtle development underexamined. Watch for essays that spend three paragraphs on the villain and a single sentence on the narrator who carries most of the book.

A third error is penalizing unconventional interpretations simply because they differ from your own reading. If a student argues that Shelter is less a monster than a product of institutional neglect, check whether the evidence supports the claim before deciding. Good marking asks whether an argument is well supported rather than whether it matches the teacher's preferred answer. Building that principle into the rubric protects students who think independently.

Using AI to Apply the Rubric Consistently

Once the rubric is written, an AI grading tool can apply it to every essay in the same way, which removes the fatigue effect that creeps in during long marking sessions. It can highlight where a student shifts from analysis back to retelling and suggest the specific quotation that could be developed further. Teachers still review the results, but the first draft of the feedback already reflects the department's criteria.

The payoff compounds across a unit. Students see the same criteria on every piece of feedback, so they begin to internalize what a strong character analysis requires. By the time they write the final essay, many have absorbed the habit of tying a quotation to an effect on the reader. A clear rubric, applied consistently, teaches as much as it measures.