Evidence use is one of the most reliable indicators of a student's reading comprehension, yet it often receives only a passing comment in feedback. In essays on When the Legends Die, students may quote a long passage about Tom's time in the mountains but fail to explain which part actually matters. Others paraphrase accurately but offer no source detail, leaving the reader unable to verify the claim. A grading approach that treats evidence as its own skill produces better writing and clearer feedback.

Start by deciding what counts as strong evidence for the assignment. Some teachers require direct quotations, while others accept specific paraphrases of scenes, particularly with younger readers or students who are still building confidence. Whichever you choose, state it in the prompt so students know how to prepare. Ambiguity here is a common source of grading disputes.

The next question is relevance. A quotation can be accurate and still weak if it does not directly support the claim being made. For instance, a student arguing that Tom grows more guarded may choose a scene where he simply speaks to another character rather than one showing his reluctance to trust. Teaching students to test each quotation against their claim improves selection more than any rule about length.

Separate Selection From Integration

Selection is about choosing the right moment, while integration is about weaving it into the student's own sentence. A student who drops a quotation into the paper with no introduction forces the reader to guess its relevance. Strong integration includes a short lead-in that identifies the scene, the quotation or paraphrase itself, and a sentence that explains its significance. Scoring these steps separately helps students see which one they need to practice.

• Chooses a quotation that directly supports the claim
• Introduces the quotation with enough context for the reader
• Keeps quoted material brief and focused on the key phrase
• Explains the connection between evidence and claim in the student's own words
• Cites the source consistently according to class expectations

Give Feedback That Models the Fix

Telling a student to "explain your evidence" often yields little improvement because the instruction is too abstract. A more effective comment demonstrates the pattern by suggesting a sentence starter, such as "This moment shows that Tom..." followed by a prompt to finish it. Even a short model sentence can unlock a writer who has been stuck. AI tools can generate these suggestions tailored to each paragraph, which is hard to do by hand at scale.

It also helps to praise strong evidence use explicitly. When a student selects a precise phrase and explains it well, point out what worked so the student can repeat it. Positive feedback of this kind is more specific than a general compliment and builds confidence. Over time students begin to recognize what good evidence handling looks like.

Address Over-Quoting and Under-Quoting

Some students pad their essays with large block quotations because they believe more text means more support. Others avoid quoting altogether and rely on loose memory of the story, which weakens credibility. A rubric can define a target range, such as one short quotation or paraphrase per body paragraph, and describe how to adjust when evidence is too thick or too thin. Clear guidance prevents guesswork.

When using AI feedback, check how the tool handles quotation accuracy. It should flag quotations that appear to be misremembered or attributed to the wrong scene, though the teacher remains the final authority on the text. Students benefit from learning that accuracy matters, since small errors can undermine an otherwise thoughtful argument. A quick verification habit saves them from careless mistakes.

Teach Evidence Skills Before the Essay

Grading evidence fairly requires that students have practiced it. Before the final essay, assign short paragraph tasks in which students pair one claim with one piece of evidence and a sentence of explanation. These low-stakes exercises are quick to review and reveal who needs support before the larger grade is at stake. They also reduce anxiety because students get feedback while there is still time to apply it.

Revisit the best examples from the class, with permission and without names, as models during a mini-lesson. Seeing a peer's work labeled and explained is often more convincing than a teacher's example. Collect these samples over the years to build a library you can reuse. Over time, the rubric, the models, and the feedback all reinforce the same expectations.