Every writing teacher knows the value of showing students what strong writing looks like. The challenge is getting students to learn from the model rather than copy it. A student who reads a mentor text and then produces an essay with the same opening structure, the same transition words, and a suspiciously similar argument has not learned to write. They have learned to mimic. The goal of mentor text instruction is to help students identify the craft decisions a writer made, understand why those decisions are effective, and then apply the same principles to their own, entirely different, writing. That is a three-step process: read like a reader, analyze like a writer, and apply like a craftsperson.

A 2026 evidence-based practices brief on writing instruction in middle and high school found that reviewing and analyzing models of effective writing helps students internalize features of high-quality text, such as structure, tone, and use of evidence, and apply them in their own work. The key word is 'internalize.' Students who are simply told to 'look at the model and do something similar' will imitate surface features. Students who are guided through a structured analysis of what the model does and why it works will internalize principles they can deploy in novel situations. The difference is in the quality of the analysis activity, not the quality of the mentor text itself.

The most effective mentor text activities use a layered reading approach. The first reading is for content: what is this piece about? What is the writer's main point or story? The second reading is for structure: how is this piece organized? Why did the writer put the sections in this order? The third reading is for craft: what specific language choices, sentence structures, or rhetorical moves does the writer use to create an effect? Each layer builds on the previous one, and by the third reading, students are engaging with the text as writers rather than as readers. This is the reading-writing connection that research consistently identifies as one of the strongest levers for writing improvement.

Choosing the right mentor text matters. The best mentor texts for classroom use are short (one to three pages), accessible in language, clearly structured, and demonstrative of two or three craft moves you plan to teach. Published essays, op-eds, narrative nonfiction, and strong student work from previous years all make effective mentors. Avoid using texts that are so sophisticated that students cannot see the underlying structure, or so simple that there is nothing to analyze. The sweet spot is writing that students can read independently but that contains craft decisions they would not have made on their own.

A Three-Read Mentor Text Protocol

This protocol works for any essay type and any grade level. The three reads can be done in a single class period (15 minutes each) or spread across two or three days depending on the text's complexity.

• Read 1 (Content): Students read for understanding. What is the main idea or story? What evidence or details does the writer use? Brief discussion or written response.
• Read 2 (Structure): Students annotate the text for organizational features: thesis location, topic sentences, transition strategies, introduction and conclusion techniques. Class discussion comparing findings.
• Read 3 (Craft): Students identify two or three specific craft moves they want to try in their own writing. They mark the move in the text and write a sentence explaining what it does and why it works.
• Application: Before drafting their own essay, students write a 'craft plan' identifying which techniques from the mentor text they intend to use and where in their essay they will try them.
• Reflection: After drafting, students highlight the places where they used mentor text techniques and assess whether the techniques worked in their own writing.

Connecting Mentor Texts to Grading

Mentor texts also serve a grading function: they calibrate student expectations. When students analyze a strong essay against the rubric before writing their own, they develop a concrete understanding of what each performance level looks like. This reduces the 'I did not know what you wanted' response to grades and makes rubric criteria tangible rather than abstract. A student who has annotated a model essay for thesis quality, evidence use, and analysis depth has a reference point for their own work that no amount of rubric language alone can provide.

For teachers using AI grading tools, mentor texts serve an additional function: they provide the exemplars against which you can calibrate the AI's scoring. Running a mentor text through the AI tool with your rubric shows you whether the AI assigns the scores you would assign. If it does, you have a well-calibrated system. If it does not, the discrepancy tells you where the rubric language needs tightening. This calibration step takes five minutes and significantly improves the accuracy of AI-assisted scoring for the rest of the semester.