Poetry essays are notoriously hard to grade consistently because two teachers can read the same paragraph and disagree about its quality. A good rubric reduces that disagreement by naming exactly what counts as strong work. For a collection like Carol Ann Duffy's Selected Poems, the rubric also needs to reflect how her poems operate, through speakers, irony, and compressed imagery. Without that fit, a generic literature rubric tends to reward length and confidence over real insight.

Start by limiting the number of criteria, because rubrics with ten rows are rarely used carefully. Four or five categories cover most of what matters, such as argument, use of evidence, analysis of language and form, awareness of context, and written expression. Each category should describe observable behavior at every level, so a student can see what moves a response from adequate to strong. If a descriptor cannot be checked against the essay text, it is probably too vague.

Weighting deserves attention too. In most classrooms, analysis of language and form should carry more weight than context, since Duffy's poems reward attention to how meaning is made. If context is worth the same as analysis, students tend to spend half their essay on biography and the Poet Laureate years. Making the weighting explicit tells students where to invest their effort.

Write descriptors that distinguish levels clearly

The hardest part of rubric design is the middle of the scale, where descriptors often collapse into words like "some" and "sometimes". Instead, describe what the student actually does. A developing essay might identify techniques correctly but explain their effect in a single general phrase, while a proficient essay connects the technique to the speaker's attitude and the poem's wider concerns. This level of specificity makes marking faster and moderation easier.

• Argument: a clear, arguable claim that develops across paragraphs
• Evidence: short, well-chosen quotations embedded in the sentence
• Analysis: explanation of how language, imagery, and structure shape meaning
• Context: relevant ideas used to explain the poem, not replace analysis
• Expression: controlled, precise academic writing with minimal errors

Test the rubric on sample responses

Before using any rubric on a full class, grade three or four sample essays, ideally ones that represent different quality levels. Notice where you hesitate, because hesitation usually signals an unclear descriptor or overlapping categories. If you find yourself giving the same essay a different mark depending on the day, tighten the language until the decision feels straightforward. This small investment prevents far bigger problems when you grade a full set.

If you work in a department, ask a colleague to grade the same samples independently and compare results. Differences of more than one band on any criterion point to wording that needs revision. These conversations are often the most valuable part of rubric design, because they surface assumptions about what good Duffy analysis looks like. The final document then reflects shared expectations rather than one teacher's habits.

Share the rubric with students early

A rubric hidden until marking day does little to improve writing. Introduce it at the start of the unit and use it to analyze a model paragraph on a poem such as "Havisham", asking students to identify which descriptors it meets. This builds the habit of self-assessment and gives them a shared vocabulary for discussing quality. Students who understand the criteria tend to ask better questions during drafting.

You can also ask students to annotate their own drafts against the rubric before submitting. Even a quick pass, marking where they think they show analysis of form or use evidence well, exposes gaps they might otherwise miss. When they receive feedback, the language matches what they have already been thinking about. That continuity makes revision feel purposeful instead of arbitrary.

Apply the rubric consistently at scale

Consistency is where a rubric proves its value, especially across large classes or multiple sections. Tools like GraideMind can apply the same written criteria to every response, flagging where an essay meets or misses a descriptor and drafting comments tied to that language. A teacher then reviews and adjusts, which keeps the final judgment human while removing much of the repetitive work. The rubric becomes a living instrument rather than a document in a drawer.

Revisit the rubric at the end of each unit and note which rows caused confusion or disagreement. Perhaps the context criterion rewarded name-dropping, or the expression row was too harsh on students writing in a second language. Small revisions each year lead to a sharper, fairer instrument. Over time, your Duffy rubric will reflect what your students actually need to learn.