NURS 766 Week 4 Written Assignment 4: Developing a Rubric Example

An Analytic Rubric for a Heart Failure Discharge Teaching Plan: Criteria, Levels, Validity and Rater Agreement

The Assignment and Its Objective

In the heart failure unit, students write a discharge teaching plan for an assigned patient: Mr. Duncan, 68, newly diagnosed with a weak left ventricle (ejection fraction of 30%), who reads at about a sixth-grade level and whose daughter can visit only on weekends. The plan measures the unit's highest objective: develop a discharge teaching plan on daily weights, sodium restriction and symptom reporting for a patient with heart failure. McDonald (2018) recommends rubrics for assignments like this one, in which students produce a complex product that no single correct answer can capture, and stresses that the rubric must be built from the objective rather than from a general sense of what good work looks like.

Analytic or Holistic

A holistic rubric would give the plan one overall score. It is quick to use, but it hides why a plan scored as it did, and two graders can give the same score for different reasons. An analytic rubric rates every criterion on its own scale and sums the results; grading slows down, but students learn exactly where they fell short and faculty can see which parts of the objective the class as a whole missed. Brookhart (2013) recommends analytic rubrics when feedback is a main purpose, and since this assignment comes midway through the course, feedback is the point. The rubric below is analytic.

Choosing and Weighting Criteria

Five criteria come directly from the objective and the patient. Three cover the required content: daily weights, sodium restriction and symptom reporting. One covers fit to the learner, since the patient's limited health literacy should change how the plan is written. One covers evaluation of learning, because a plan without a way to check understanding cannot be judged effective. Writing mechanics are not a separate criterion; clarity counts within the fit criterion, where it affects the patient.

Weights reflect risk. Symptom reporting carries the most weight because a patient who does not know when to call will return to the emergency department or worse. Daily weights and sodium restriction come next. Fit to the learner and evaluation of learning carry less weight but cannot be ignored. The total is 50 points.

The Rubric

Each descriptor names something a reader can find in the plan, such as a threshold, a method or a teach-back question, rather than a judgment such as "good" or "thorough." The levels differ in what is present, not in adverbs.

Writing the Descriptors

The descriptors were drafted from twelve teaching plans submitted in the previous term, with names removed. Reading them side by side showed where the real differences lay. The strongest plans gave Mr. Duncan a specific weight gain that should prompt a call and told him whom to call; average plans told him to "watch for swelling" without a threshold; weak plans left out symptom reporting altogether or gave advice that could delay care, such as waiting a week to see whether swelling improves. Those observed differences became the four levels for the symptom reporting criterion. The same process shaped the other criteria: the plans showed that students almost always mentioned sodium but rarely gave a limit or practical strategies, so the sodium descriptors separate those elements.

Writing from real work also exposed a problem with the old rubric, which had a criterion called "organization" that rewarded headings and bullet points. Several well-organized plans contained unsafe advice and scored well anyway. Dropping that criterion and weighting by patient risk keeps formatting from outscoring safety. Brookhart (2013) warns against criteria that describe the surface features of a product rather than the learning it should show, and the organization criterion was a clear case.

Content Validity

Before use, three reviewers will rate each criterion and descriptor for relevance to the objective on a four-point scale: a heart failure nurse navigator from the clinical partner, a faculty member who teaches the course and a faculty member outside the course. Criteria rated low by two reviewers will be revised. The navigator's review matters most for the content criteria, since the navigator knows what patients going home actually need to hear.

Reliability

A rubric that two graders apply differently does not measure the objective, however well it is written. Diez-Fernandez et al. (2026) tested an eight-item rubric for a simulated nursing procedure with 186 students and found interrater agreement ranging from moderate to substantial, with kappa values from 0.418 to 0.796, and adequate internal consistency, with agreement weakest on items requiring judgment about technique. The lesson for this rubric is that the criteria requiring judgment, fit to the learner and evaluation, will need the most attention. Before the term, the two course faculty will score five sample plans independently, compare scores, discuss differences and revise descriptors, then score five more and calculate Cohen's kappa for each criterion. A kappa below 0.6 on any criterion will send it back for revision.

Designing Around Generative AI

A teaching plan is the kind of product a generative AI tool can draft in seconds. Salamonson et al. (2026) argue that rules forbidding AI use are largely unenforceable and that assessments should instead be changed structurally, in how they work, to keep their validity. The evaluation criterion has been redesigned in that spirit: students now record a five-minute teach-back session with a classmate playing the patient, and the rubric scores the questions they actually ask and how they respond to a wrong answer. Drury et al. (2026) caution that AI-assisted feedback, though efficient for large classes, struggles to capture nuance and context-dependent competencies, and they propose keeping humans in the loop. Faculty will continue to grade this assignment themselves. The rubric can still be offered to students as a self-check before they submit, which gives them the benefit of rubric-aligned feedback without handing the grading decision to a tool. Students who score their own draft and then compare it with the faculty score also learn how the descriptors are read, which tends to make the next submission stronger.

Using and Revising the Rubric

The rubric is posted alongside the assignment instructions, and one class session will include grading a sample plan together, so they see how the descriptors work. After the first term, the faculty will review the score distribution for each criterion; a criterion on which nearly every student scores exemplary may be too easy, and one on which most score developing may signal a teaching gap rather than a rubric problem. A rubric is finished only when students can use it to improve their work before it is graded.

Conclusion

The rubric turns the unit's highest objective into five weighted criteria with descriptors a reader can verify, and it comes with a plan to check validity with experts and reliability with an agreement statistic. Its structural change for generative AI keeps the evaluation criterion tied to what the student actually does. The same approach, criteria from objectives and descriptors that name observable qualities, will shape the clinical evaluation tool at the end of the course, where the stakes for students are higher still and the need for agreement between raters is greater.

References

Brookhart, S. M. (2013). How to create and use rubrics for formative assessment and grading. ASCD.

Diez-Fernandez, J. M., Fernandez-Villa, T., Rodriguez-Badiola, A., Mauriz, E., Crespo-Martinez, C., & Vazquez-Casares, A. (2026). Validity and internal consistency of a rubric for cervical collar placement in nursing students. Nursing Reports, 16(7), 250. https://doi.org/10.3390/nursrep16070250

Drury, P., Chan, A., Bowdler, S., Kinghorn, G., & Curtis, E. (2026). AI-assisted feedback in nursing education: A human-in-the-loop assessment framework for ethical and pedagogically sound practice. Nurse Education Today, 168, 107411. https://doi.org/10.1016/j.nedt.2026.107411

McDonald, M. E. (2018). The nurse educator's guide to assessing learning outcomes (4th ed.). Jones & Bartlett Learning.

Salamonson, Y., Irwin, P., Kornhaber, R., Kirk, D., McGough, S., Vakavosaki, K., Ramjan, L., Bloomfield, J., Matheson, A., & Greenwood, M. (2026). Refining assessment design and practice in nursing and midwifery education in the era of generative AI: A discussion paper. Journal of Clinical Nursing. Advance online publication. https://doi.org/10.1111/jocn.70484