The sixth stage of NR-670 is where the project has to become checkable. The written work is a measurement and evaluation plan: which outcome, process and balancing measures you will use, how each is defined and collected, from what baseline, at what interval, and what result would count as improvement rather than noise. Written well, it is what allows the final paper to say anything at all about whether the work mattered. Your section may print this as NR 670 or NR670; it is the same course. Chamberlain publishes no syllabi outside Canvas. The placement here is our teaching judgment from the course's catalog arc; your section's rubric decides what your week actually asks.
What a measurement plan has to guarantee
An informatics-minded nurse leader building a capstone around discharge instruction completeness discovered her problem in an audit spreadsheet, not in the literature. Two auditors had reviewed the same twelve records and disagreed on five of them, because complete had never been defined. Writing an operational definition, four named fields present and legible, took her one paragraph and turned an unusable measure into one that two people could apply identically. That paragraph is the heart of this stage.
A measurement plan has three families of measure and a capstone should carry at least two of them. Outcome measures say whether the thing you care about changed: readmissions, discrepancies, delays, harm events. Process measures say whether the change actually happened: how often the new step was performed as designed. Balancing measures say whether you broke something else: added minutes at handoff, a workaround appearing elsewhere, a satisfaction score sliding. Students almost always write outcome measures and almost always forget balancing ones, and the omission is visible from across the room to anyone who has run improvement work.
The second requirement is operational definition. Every measure needs a numerator, a denominator, an inclusion rule, an exclusion rule, a data source, a collection interval and a named collector. A measure that cannot be handed to a colleague with those seven items attached is not a measure, it is a hope. This is also where the small baseline you established earlier gets its second use: the same operational definition has to apply to the before and after, or the comparison is meaningless.
The practicum boundary holds here in a specific way that is easy to get wrong. Your hours, your activity log and any evaluation your mentor completes remain your own verified record, never drafted, reconstructed or estimated with help, and the same principle extends to data: results are collected, not composed. A measurement plan is a written design for how findings will be produced; it is never a place to write findings that did not occur. Where the plan describes real records, everything is de-identified and reported as aggregate counts.
The NR-670 Week 6 method, step by step
Six moves that make a change checkable by someone other than its author.
-
Return to the unit you fixed in the problem statement
The primary outcome measure should be the same quantity you used to establish the gap, defined identically. Changing the unit between stages is the most common reason a capstone cannot demonstrate its own result.
-
Write an operational definition a stranger could apply
Numerator, denominator, inclusions, exclusions, source, interval, collector. Then test it: hand the definition and five records to a colleague and see whether their count matches yours. Disagreement now is cheap; disagreement in week eight is fatal.
-
Add a process measure tied to the behavior you designed
If the change is a four-field transfer note completed before departure, the process measure is the proportion of transfers where it was. Process measures move first and tell you whether a flat outcome means the change failed or was never delivered.
-
Name at least one balancing measure
Something that would get worse if your change is more burdensome than it is worth: added time at a step, an alternative form left blank, a downstream complaint. Choosing it yourself is a stronger signal than any outcome you could report.
-
Decide how the data will be displayed before it exists
Small samples over a short practicum are better shown as counts over time or as a run of consecutive periods than as a single before and after pair. Say what display you will use and why, since a display chosen after the fact invites the suspicion that it was chosen to flatter.
-
State in advance what would count as improvement
A threshold, a direction sustained over a stated number of periods, or a defined shift. Committing before the data arrives is what separates evaluation from advocacy, and it is a competency the rubric language at this level usually reaches for.
A layout and word budget for a measurement plan
Our frame for an evaluation plan, sized for roughly 1,200 to 1,500 words plus the measure table. It is our own outline rather than anything the university issues, and your section's rubric outranks it wherever the two disagree. Scale proportionally if your assigned length differs.
| Section | What belongs in it | Word target |
|---|---|---|
| What is being evaluated | The change as implemented, the aim restated in measurable terms, and the period the evaluation covers. | 130 to 160 |
| Outcome measure | Full operational definition, source, interval, collector, and its link to the baseline you established earlier. | 250 to 300 |
| Process measure | The designed behavior expressed as a proportion, with the same seven definitional elements attached. | 200 to 250 |
| Balancing measure | What would deteriorate if the change costs more than it returns, and how you would detect it. | 170 to 210 |
| Data handling and display | Collection route, storage, de-identification, who has access, and the chart or table the results will appear in. | 200 to 250 |
| Decision rules and limitations | What counts as improvement, what would trigger revision, and the threats to interpreting a change as caused by your work. | 230 to 290 |
Evidence craft for evaluation writing
Attribute your measurement framework. If your measure set follows a published improvement or evaluation model, name it with its source and year and use its categories. Faculty at capstone level read for whether the structure was borrowed deliberately or assembled ad hoc.
Never report a rate without its base. Nine of 34 is a finding; twenty-six percent from an unstated denominator is a decoration. In short practicums the denominators are small enough that hiding them changes how a reader should weigh every claim you make.
Use tentative causal language and mean it. An uncontrolled before and after in one unit cannot establish that your change caused the difference. Write associated with, coincided with, or followed the introduction of, and then name the other things that changed in the same period. Precision here reads as competence, not as weakness.
Say who collects, and whether they are invested in the result. A process measure audited by the person who designed the change is a legitimate practicum arrangement and a known source of bias. Name it as a limitation and, where possible, describe the check you built in, such as a second reviewer on a sample of records.
Write the de-identification step into the plan, not just the paper. State that identifiers are removed at the point of collection, that only aggregate counts leave the system, and that no case-level detail appears in the academic document. That sentence protects you and demonstrates the data governance judgment a leadership rubric is looking for.
Five mistakes that cost points in this week's territory
- Measuring satisfaction with the project. Whether staff liked the change is not whether it worked, and it is the easiest measure to collect for exactly that reason.
- No process measure. Without one, a flat outcome cannot be told apart from a change that was never actually performed.
- Definitions that drift between baseline and follow-up. If complete meant something different in week three, the comparison in week eight is not a comparison.
- Causal claims from an uncontrolled design. Reduced belongs to controlled comparisons; a single-unit before and after supports followed and coincided with.
- Success criteria written after the data arrives. Deciding what counts as improvement once you can see the numbers is the definition of a biased evaluation.
Before you submit
- The primary measure uses the same unit as the original problem statement
- Every measure has numerator, denominator, inclusions, exclusions, source, interval and collector
- At least one process measure and one balancing measure appear
- The display format is named before any data is discussed
- Success criteria are stated in advance, with a number or a sustained direction
- Causal language matches the strength of the design
- De-identification is described as part of the collection process
Writing the NR-670 evaluation plan?
Send the rubric and your baseline out of Canvas. A premium original draft comes back in 24 to 48 hours with operational definitions a stranger could apply, a balancing measure chosen, and causal language matched to the design, and revisions run until the grade lands.