If your DNP methods section says "nurses will be educated and data will be analyzed", a committee cannot tell who does what. This guide is for doctoral students who have a problem and an evidence-supported intervention and now need a design, sample, protocol, measures and analysis plan. The eight decisions are design, setting, sample, intervention protocol, measures, data collection, analysis and fidelity.
Quick answer. Good project methods answer six questions: what design, where, with whom, doing exactly what, measured how, and analyzed how. If another clinician cannot repeat your project from the description, the section is not finished.
Most DNP projects are quality improvement or evidence-based practice implementation projects, not experiments. Your design should be justified by your aim and by what the site can support. Ask your chair which designs your program accepts before you invest in a plan.
| Design | Fits when | Strength | Limitation to state |
|---|---|---|---|
| Pre and post (one group) | You have a defined baseline and one site | Simple, feasible | Cannot separate your change from other changes over time |
| Repeated measures over time (run chart style) | Data can be collected weekly or monthly | Shows trends and variation | Needs enough data points before and after |
| Comparison with a similar unit | Another unit will share data | Adds a reference for outside influences | Units may differ in ways that matter |
| Pilot with feasibility outcomes | You are testing whether a change can be delivered at all | Low risk, informs scale-up | Not designed to prove effectiveness |
| Program evaluation | A program already runs and needs assessment | Fits real-world delivery | Less control over how the program was implemented |
Every design has weak points. Naming them in the methods section, and saying what you will do about them, is a sign of scholarly maturity. It also pre-empts the questions a reviewer would otherwise ask at your defense.
| Threat | How it shows up in a practice project | What you can do |
|---|---|---|
| Other changes over time | A new policy or seasonal patient mix changes the outcome | Track known events on the run chart and discuss them |
| Staff turnover | People delivering the intervention differ from those trained | Refresher training and a quick guide for new staff |
| Measurement change | A record template or definition is updated mid-project | Freeze definitions or note the date of any change |
| Observation effect | Staff behave differently because they know they are audited | Use routine data where possible and acknowledge the effect |
| Regression to the mean | An unusually bad baseline month naturally improves | Use a baseline long enough to show typical variation |
Describe the setting in enough detail that a reader can judge whether your results might apply elsewhere. Context is not decoration. It shapes what you can do and how well it will work.
| Feature | Why it matters |
|---|---|
| Type and size of unit or clinic | Sets the scale of the change and the sample |
| Staffing model and shifts | Affects who can deliver the intervention and when |
| Patient population served | Shapes inclusion, materials and outcomes |
| Existing electronic tools and templates | Determines what can be built and what data exist |
| Competing initiatives | Risks to uptake and to attribution |
Keep facility details general enough to protect anonymity if your program requires it, for example "a mid-sized adult medical unit in an urban teaching hospital".
State who is in the project and how they are chosen. In improvement work the "sample" may be patient records, patients, staff or all of these, so specify each.
Formal power calculations are more common in research than in improvement projects. If your program asks for one, a statistician or your chair can help; our biostatistics guide explains the ideas. If it does not, justify the size by the volume available and the number of data points you need to see a pattern.
When staff are part of the project, describe how they are told about it and how they can decline. Voluntary participation is easier to protect when you do not supervise the people you invite. If you hold a management role, agree with your chair who will make the invitation.
Write the intervention so someone else could deliver it. Think of it as a recipe, not a description of an idea.
| Question | What to write |
|---|---|
| Why | The evidence-based rationale, in one or two sentences with citations |
| What | The materials, tools and steps, in order |
| Who | The role delivering each step and their training |
| How | The mode, such as in person, by phone or in the electronic record |
| Where and when | The location, timing, frequency and duration |
| Tailoring | What can be adapted, and what must stay the same |
The TIDieR checklist, available through the EQUATOR Network, is a well-known model for describing interventions completely, and you can borrow its logic even if you are not required to use it.
Send your aims, setting and intervention outline with your brief, and we can draft or edit the methodology so each choice is justified. The price is shown before you pay, and every delivered paper includes 14 days of free revisions.
A measure is only useful if two people would count it the same way. Write an operational definition for each, including what counts and what does not.
| Measure | Type | Operational definition (illustrative) | Source | Frequency | Owner |
|---|---|---|---|---|---|
| Structured discharge teaching completed | Process | Proportion of eligible discharges with the teaching template fully documented | Chart audit | Weekly | Student with unit analyst |
| Thirty-day return to hospital | Outcome | Proportion of index discharges with an unplanned return within 30 days, per the organization's definition | Quality report | Monthly | Quality analyst |
| Time added to discharge | Balancing | Self-reported minutes added by the new step | Short staff form | Every two weeks | Student |
| Staff acceptability | Implementation | Mean score on a brief acceptability tool | Anonymous survey | End of project | Student |
Use validated tools where they exist. If you use a published instrument, cite its development and reliability evidence, and ask for permission when required. If you create a tool, say so, and describe how you checked that staff understood it.
More measures do not make a stronger project. A small set that you can collect reliably is better than a long list you cannot sustain. As a general habit, choose one primary outcome or process measure, a small number of supporting measures, and one balancing measure. If your chair recommends a different mix, follow that advice.
Explain who collects each item, when, and where it is stored. This is often the part committees find thin.
State whether data are de-identified, where files are kept, who can access them and when they will be destroyed. Follow your institution's rules and those of the site. The IRB and research ethics guide shows how to phrase this for an ethics application.
Write the analysis plan before collecting data. It tells reviewers you will not go hunting for a result. Match each method to a measure and to a data type.
| Data type and question | Typical approach | Notes |
|---|---|---|
| Repeated proportions over time (for example weekly completion rate) | Run chart or control chart with median or mean | Needs several points before and after |
| Paired continuous scores, same people before and after | Paired t-test, or Wilcoxon signed-rank if not approximately normal | Check the assumptions |
| Two independent groups, continuous | Independent t-test, or Mann-Whitney U | Use when pre and post samples are different patients |
| Two proportions, independent groups | Chi-square, or Fisher exact for small counts | Report counts as well as percentages |
| Paired binary outcome | McNemar test | Same people, yes or no at two times |
| Open-ended feedback | Content analysis with a simple coding frame | Show how codes were developed |
Report effect size or the size of the change alongside any p-value, and remember that a small project may show a meaningful change that is not statistically significant. Say what change the site would consider worthwhile before you look at results. Your outcomes chapter, described in the outcomes evaluation guide, picks up from here.
Run charts plot a measure over time with a median line. Improvement guidance from groups such as the Institute for Healthcare Improvement describes a few common rules for judging whether a pattern is more than chance. Confirm which rules your program follows, and apply them consistently.
| Signal | Common rule of thumb | What it suggests |
|---|---|---|
| Shift | Six or more consecutive points all above or all below the median | A sustained change in the process |
| Trend | Five or more consecutive points all rising or all falling | A gradual change in one direction |
| Points on the median | Do not count them when checking a shift | Avoids over-reading the pattern |
| Unusual point | A single point far from the rest | Check for a data error or a special cause |
Fidelity is how closely the intervention was delivered as designed. Without it, you cannot tell whether a weak result means a weak intervention or a weak delivery.
Proctor and colleagues described a set of implementation outcomes that many projects borrow: acceptability, adoption, appropriateness, feasibility, fidelity, cost, penetration and sustainability. You do not need all of them. Pick two or three that fit your question.
| Implementation outcome | Simple way to capture it |
|---|---|
| Acceptability | Short staff survey or huddle feedback |
| Adoption | Share of eligible staff or visits using the tool |
| Fidelity | Checklist of protocol steps, completed during audits |
| Feasibility | Time per use, barriers logged |
| Penetration | Proportion of eligible patients reached |
Real projects change. Record each adaptation, the date, the reason and who approved it. That log turns a messy implementation into useful evidence.
A short pilot on a few patients or one shift finds problems cheaply. Test the training, the data form and the workflow, then adjust before the full start. Many students describe this as a small Plan-Do-Study-Act cycle.
The Institute for Healthcare Improvement provides free guidance on this cycle and on the Model for Improvement.
| Feedback | What it means | Fix |
|---|---|---|
| "I could not repeat this." | The protocol lacks detail on who, when and how. | Add the protocol template and a one-page tool as an appendix |
| "Your measure is unclear." | No operational definition. | Add numerator, denominator, source and timing |
| "Why this test?" | The statistic does not match the data type. | Use the matching table and justify the choice in one sentence |
| "What about bias?" | Threats to validity are not discussed. | Add the threats table and your responses |
The layout below is invented to show how the subsections fit together. It is not real data.
| Subsection | Illustrative entry |
|---|---|
| Design | Quality improvement project with a pre-implementation baseline and weekly monitoring. |
| Setting | One adult cardiac step-down unit in a community hospital. |
| Sample | All adult patients discharged home during baseline and implementation; unit nurses as staff participants. |
| Intervention | A teach-back discharge protocol delivered by the discharging nurse using a one-page tool. |
| Measures | Process: teaching documented. Outcome: return within 30 days. Balancing: minutes added. |
| Analysis | Run chart for the process measure, descriptive comparison for the outcome, thematic summary of staff comments. |
| Fidelity | Weekly audit of five records with a step checklist and an adaptation log. |
| Weak | Stronger |
|---|---|
| Nurses will be educated about the tool. | All unit nurses will complete a 20-minute session led by the project lead, with a one-page guide and a supervised first use. |
| Data will be collected and analyzed. | The unit analyst will supply weekly de-identified counts; the student will plot them on a run chart and compare the baseline and implementation medians. |
| The project will improve outcomes. | The primary outcome is the proportion of eligible discharges with complete teaching documentation. |
| Fidelity will be monitored. | Five random records per week will be audited against a six-step checklist, and results will be shared at the weekly huddle. |
Illustrative example, not a real client. This short story is invented to show the pattern, and it contains no real people or numbers.
The problem. A student's methods section said, "Nurses will be educated and data will be analyzed."
The tension. The committee could not tell who would do what, and the site's quality analyst asked how each measure would be counted.
The turn. She wrote the intervention as a step-by-step protocol, built a measure specification table and tested her data form on a few records before launch.
The proof. The analyst confirmed she could supply the data in that format, and the committee approved the section without further changes.
The payoff. Implementation ran without a redesign, and a mid-project change was logged and explained instead of hidden.
Not necessarily. Many practice projects rely on a baseline period. State what that design can and cannot show, and avoid causal language that it does not support.
They can add value, especially for barriers and staff experience. Keep them proportionate to the project and explain how you will analyze the responses.
Say so and focus on descriptive results, run charts and practical significance. Avoid strong claims from tests with very few observations.
That depends on your institution and site. Ask early, and let the review body make the determination.
SQUIRE 2.0 is widely used for improvement reports, and other guidelines exist for other designs. Ask your chair which one to use.
Strong methodology is specific, consistent and honest about its limits. When each measure, step and analysis can be traced back to an aim, the committee can focus on your ideas rather than on gaps in your plan.
Want your methods section drafted or edited? Get my instant quote. The price is shown before you pay, every delivered paper includes 14 days of free revisions, and refund terms are on the money-back guarantee page. Please use any model paper in line with your institution's academic-integrity policy.