NR-709A · Week 4 of 8 · Displaying change over time

NR-709A Week 4 Displaying Change Over Time: How to Write It

The short answer

Halfway through an evaluation term the numbers have to become something a reader can see, and the choice of display is an analytic decision rather than a cosmetic one. Improvement work is about change over time, which means a time series with the implementation annotated on it almost always tells the truth better than two bars comparing a before average with an after average. This stage is about building honest displays, labeling them properly, and writing the prose that says what each one shows. Your section may print this as NR 709A or NR709A; 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.

NR 709A Week 4 grading scale at Chamberlain, the criterion levels this assessment is scored on, from Chamberlain Tutors
How Chamberlain grades NR 709A Week 4, visualized by Chamberlain Tutors.

What NR-709A Week 4 asks for

What can two bars hide? A blood pressure follow-up program running out of a food pantry site reported its result as a pair of columns: a baseline average of thirty-eight percent and a post-implementation average of fifty-one. The improvement looked clean and the committee was pleased. When somebody finally plotted the same data by month, the picture was entirely different. The measure had been climbing steadily for five months before implementation, had dipped during the two weeks the outreach worker was out, and had flattened at forty-nine percent well before the project's final component went live. The average comparison was arithmetically correct and told a story the data did not support.

That is why improvement evaluation prefers time-ordered display. A run of points across months lets a reader see the level before the change, the trend it was already on, what happened at the moment of implementation, and whether the new level held. It also makes the ordinary variation visible, which is the context that decides whether a shift means anything. Nothing in a doctoral evaluation improves a paper as reliably as a well-built time series with the implementation date marked and the interruptions annotated.

Expect a deliverable that presents results, often with required figures or tables, sometimes accompanied by a narrative interpretation. Two disciplines apply. First, every figure needs a caption that lets it stand alone: what is plotted, for which population, over what period, from what source. Second, the prose must state what the figure shows rather than leaving the reader to infer it, because a rubric row that asks you to analyze results cannot be satisfied by a graphic.

The boundary that governs this manual. Practicum hours, logs, encounter counts, site documentation, preceptor evaluations and signatures belong to you and your site, are never drafted, reconstructed or estimated with outside help, and no tutor performs or documents clinical activity on anyone's behalf. What is supportable is the written and analytic layer around real work you did: choosing an honest display, labeling it, and writing the interpretation. Any patient detail is de-identified before it appears, small cell counts are handled so that no individual becomes identifiable, and the site is described by type rather than by name.

The NR-709A Week 4 method, step by step

Seven moves for building displays that tell the truth about your project.

  1. Plot the measure by period before you decide anything

    Monthly or weekly points across the whole window, baseline through post-implementation. Look at it before choosing any summary statistic, because the shape of the series determines which summaries are honest.

  2. Mark the implementation date on the chart

    A vertical annotation at go-live, plus markers for any component that landed later. An unannotated time series makes the reader guess where the intervention sits, and guessing readers reach their own conclusions.

  3. Annotate the interruptions you already documented

    Staff absence, a renovation, a system upgrade, a holiday period, a suspended workflow. These belong on the chart as small labels. They explain dips that would otherwise look like failure and rises that would otherwise look like success.

  4. Show the denominator alongside the rate

    A percentage from a month with eleven eligible visits deserves less weight than one from a month with a hundred and forty. Put counts in a table beneath the figure or label them on the points, so the reader can weigh each period.

  5. Use a run chart properly if you use one

    A center line drawn from the baseline period, points across time, and the standard rules for reading shifts and runs applied and cited. If you are not going to apply the rules, plot a simple annotated line and say so rather than implying an analysis you did not do.

  6. Keep every axis honest

    Start percentage axes at zero unless you have a stated reason not to, and if you truncate, say so in the caption. Consistent scales across comparable figures, no dual axes, no three-dimensional effects, no smoothing that hides the variation.

  7. Write one interpretive paragraph per figure

    What the reader should see, what it does and does not establish, and which annotation matters most. The paragraph is what the rubric scores; the figure only makes it faster to follow.

A layout and word budget for a results presentation

Our frame for the results section of an evaluation report, sized for roughly 1,100 to 1,400 words plus figures and tables. It is our own outline rather than anything the university issues, and your week's rubric outranks it wherever the two disagree.

SectionWhat belongs in itWord target
What is presented and whyThe measures shown, in what order, and the logic of that order for a reader meeting them fresh.110 to 140
Population and periods describedHow many encounters or patients fall in each period, and how comparable the periods are.150 to 190
Primary outcome over timeThe annotated series, its caption, and the paragraph saying what it shows and what it cannot.230 to 280
Process measure over timeWhether the change was actually delivered, displayed on the same time axis for comparison.190 to 230
Balancing measureWhat was watched for harm, what it did, and how confidently that can be read at this scale.150 to 190
Subgroup or site variationWhere the data allows it, how the pattern differs across groups, with counts attached.170 to 210
What the displays do not settleThe questions the figures raise and leave open, placed before the discussion section resolves anything.120 to 150

Evidence craft for presenting improvement data

Captions must let a figure travel alone. Figures get pulled into slide decks and committee packets without their paragraphs. Each caption should name the measure, the population, the period, the data source and the meaning of any annotation, so that a version separated from your report still tells the truth.

Cite the method behind any chart that implies one. Run chart rules and control chart limits come from published methodology, and using them without attribution leaves a reader unsure whether the analysis was performed or the shape merely resembles it. Name the source and the specific rules you applied.

Report counts everywhere you report proportions. Twenty-three of ninety-one eligible visits carries information that twenty-five percent does not, particularly across small monthly cells. This is also a protection: proportions from tiny denominators swing wildly and invite overinterpretation by everyone including you.

Suppress or aggregate small cells. In a small clinic or a narrow subgroup, a cell of two or three can identify a person even without names. Aggregate periods, combine categories, or suppress the cell with a note explaining the suppression rule you applied.

Present the unfavorable series with the same care as the favorable one. Use the same axis, the same period, the same level of annotation. Selective display quality is visible to any reader who is paying attention, and in a document that will circulate at your site it costs more credibility than a disappointing result ever would.

Five mistakes that cost points in this week's territory

  • Two bars where a series belongs. A before and after average conceals pre-existing trend, which is the single most common way improvement results are overread.
  • Unannotated implementation. A time series with no marked go-live leaves the central question of the paper visually unanswered.
  • Truncated axes without disclosure. A y-axis starting at thirty-five percent makes a two-point change look decisive, and disclosure is the difference between emphasis and distortion.
  • Rates with no denominators. Monthly percentages from unstable denominators produce a jagged chart that means nothing without counts.
  • Figures left to speak for themselves. Rubric rows about analysis are earned in prose; a chart with no interpretive paragraph scores as a chart.

Before you submit

  • The primary measure is displayed over time, not as two summary values
  • Implementation and any staged components are annotated with dates
  • Known interruptions appear on the chart
  • Counts accompany every proportion, by period
  • Axes are honest and any truncation is disclosed in the caption
  • Any run or control chart rules used are cited and actually applied
  • Each figure has a stand-alone caption naming measure, population, period and source
  • Every figure has an interpretive paragraph beside it
  • Small cells are aggregated or suppressed so no individual is identifiable

Building the results section for NR-709A?

Send the rubric and your measure data out of Canvas. A premium original draft comes back in 24 to 48 hours with the series annotated, denominators shown, captions written to travel alone and an interpretive paragraph beside each figure, and revisions run until the grade lands.

Questions students ask about this stage

I only have a few months of data. Is a time series still worth plotting?
Yes, with honesty about what a short series can support. Even five or six points show you whether the measure was already moving, whether the change coincided with implementation and whether the new level held for more than a moment, and all three are invisible in a before and after average. What a short series cannot support is a formal reading of shifts and runs, which needs more points than you have, so plot it as an annotated line and say plainly that the series is too short for those rules to apply. Then let the display do its honest work: a reader can see the shape, the annotation tells them when the change happened, and your paragraph says what is and is not distinguishable. Short and candid beats long and implied.
My measure went up and then came back down. How do I present that?
Exactly as it happened, and then work on explaining it, because a rise that decays is one of the most informative patterns in improvement work. Plot the whole series so the decay is visible, and annotate what changed around the point where it turned: a staffing change, the end of an active monitoring period, the departure of the person who was reminding everyone, a competing priority arriving on the unit. That pattern usually means the change depended on attention rather than on structure, which is a direct and valuable input to your sustainability section. Resist the temptation to end the chart at the peak or to report a post-implementation average that blends the rise and the decay into one flattering number, because both choices will be noticed and neither helps the site decide anything.
Table or figure for the same data?
Figure for the pattern, table for the numbers, and never both for the same content unless your rubric requires it. A time series belongs in a figure because its whole value is the shape. A comparison of specified measures across two periods, with counts, denominators and percentages, belongs in a table because a reader wants to look up values rather than estimate them from a line. Where you need both, put the series in the figure and the underlying counts in a compact table beneath it, which also solves the denominator problem cleanly. Follow your assigned style for numbering, titles and placement, since formatting compliance is usually its own scoring row, and check that every figure and table is referred to by number somewhere in your text.
Should I run a statistical test on my before and after numbers?
Usually not, and the reasons are worth stating in the paper rather than leaving implicit. A single-site improvement evaluation with data collected over consecutive time periods violates the assumptions most simple tests rest on, particularly independence, and a p value attached to a before and after comparison tends to create an impression of rigor the design cannot support. Time-ordered display with clearly marked implementation communicates more and claims less. If your rubric requires a test, or your faculty advises one, run it, report the assumptions it makes and state plainly that statistical significance in this context does not establish that your change caused the difference. That sentence is not hedging; it is the correct description of what a quality improvement evaluation can and cannot conclude.

Keep going

Online now