Skip to main content

Audience-Facing Text: What You Can and Cannot Change for A7

Where this fits

dash03 is the file you will adapt for Assignment 7. This page draws the line between audience-facing text changes (allowed for A7) and analytic changes (not allowed for A7). Knowing where that line is keeps the assignment focused and keeps your audit trail clean.

The rule, in one sentence

For Assignment 7, you may change anything the audience reads. You may not change anything the analysis computes.

Allowed changes (text-only)

These are the kinds of changes the A7 rubric rewards. They sharpen the audience experience without touching the analysis.

  • Rename a label. “Cycle” becomes “Survey wave (NHANES cycle)” so a non-expert audience knows what to read.
  • Expand a caption. A two-word caption becomes a one-sentence caption that names the statistic, source, exclusions, and limitation.
  • Add a help text line. Beside the control, a sentence that says “This dashboard summarizes adults age 20-80 from the prepared classroom dataset.”
  • Add a limitation sentence. Beside the figure, one sentence on what the artifact does not support.
  • Improve a README sentence. A README that says “Dashboard for NHANES” becomes “Descriptive dashboard of unweighted mean BMI by income group across NHANES cycles, for classroom use.”
  • Adjust a title for the audience. A neutral title becomes one that names the audience-facing question.

Not allowed for A7 (analytic-only changes)

These changes touch the analysis. They are not part of A7. If you find a real analytic issue while testing, raise it with your instructor — do not silently change it.

  • Changing the cohort filter. Switching from Age >= 20 to Age >= 18 is an analytic decision.
  • Adding or removing a survey weight. Switching from unweighted to weighted is an analytic decision.
  • Adding a regression line. Smoothing across cycles is an analytic decision (and usually inappropriate — see viz04).
  • Replacing mean() with median(). Changing the central tendency statistic is an analytic decision.
  • Changing the grouping variable. Switching from IncomeGroup to Education is an analytic decision.

If you are unsure whether a change is text or analytic, ask: “would a reader looking at the rendered HTML notice this change if I didn’t tell them?” If yes, it is probably text. If the numbers change, it is analytic.

Before / after examples

Before / after: a label

Before: selectInput("strat_var1", "Primary Grouping (Rows):", ...)

After: selectInput("strat_var1", "Compare across (rows of the chart):", ...)

The label is friendlier to a non-technical audience. The variable, choices, and computation are unchanged.

Before / after: a caption

Before: caption = "BMI by income"

After: caption = "Mean BMI by income group across five NHANES cycles. Adults age 20-80 in the prepared classroom dataset. Unweighted descriptive summary; records with missing BMI, income, or cycle are excluded."

The caption now names the statistic, source, cohort, exclusions, and disclosure. The figure itself is unchanged.

Before / after: a help text line

Before: (no help text near the control)

After: Right under the control: > This dashboard uses adults age 20-80. To see a different income group, change the value and re-render.

The audience now understands what the artifact covers and how to use it.

Before / after: a README sentence

Before: # Dashboard

After: `# NHANES Health Equity Classroom Dashboard

Descriptive, unweighted mean BMI by income group across NHANES cycles. For classroom learning; not for population inference.`

The README now names the audience and the boundary.

What to record in your A7 before-after-text.md

A7 asks you to submit one before-after text change.

Create before-after-text.md in the A7 submission folder. In the Explorer panel of your Codespace, expand assignments/assignment07-dashboard/, right-click the submission folder, choose New File, and name it before-after-text.md. Plain Markdown is fine — a short heading for each part below is enough.

For that file, paste:

  • the original text verbatim,
  • the revised text verbatim,
  • one sentence on why the change helps the audience,
  • one sentence confirming no analytic logic was changed.

That last sentence is what keeps the audit trail clean.

Where this goes next

dash05 covers how to test the starter dashboard locally, how to keep your publishing privacy-safe, and (briefly) the Shiny exemplar your instructor will demo in class.