PHLT 8076 Week 10 Misleading Health Map Analysis Example

"No Hot Spots" With Every Number Correct: How Equal Intervals Stretched by One Small Tract, a Pale Palette and a Statewide Frame Hid a County's Late-Stage Breast Cancer Pattern

Student Name

Walden University

PHLT 8076: Fundamentals of Mapping and Geographic Information Systems for Public Health

Instructor Name

Month Day, Year

"No Hot Spots" With Every Number Correct: How Equal Intervals Stretched by One Small Tract, a Pale Palette and a Statewide Frame Hid a County's Late-Stage Breast Cancer Pattern

The Map and Its Claim

The composite county's public health dashboard includes a tract map showing what share of 2018-2022 breast cancer diagnoses were late stage. Its caption reads: "No late-stage hot spots identified." The map is shaded in five classes of a pale blue palette, and a reader scanning it sees a nearly uniform county. Every value on the map is correct. Yet the same data, mapped differently in Week 4, show a clear cluster of high percentages along the county's southern edge. This analysis explains how three design choices, each defensible on its own, combined to produce a picture of calm. All figures are composite.

Choice 1: Equal Intervals Stretched by One Tract

The dashboard divides the range of tract percentages into five equal steps. The lowest tract is at 8%, but the highest is a small rural tract at 80%, where four of its five diagnoses were late stage. With the range set from 8% to 80%, each class spans about 14 points. The southern cluster, at 42% to 52%, falls into the third class, the same as many tracts in the middle of the distribution. The top two classes hold only that one small tract. In effect, a single unstable value redefined the scale for the whole county. Brewer and Pickle (2002) found that equal-interval classification performed poorly for epidemiological data with skewed distributions, and this map shows why: one extreme value compresses everything else into a few middle classes.

Choice 2: A Pale Palette

The palette runs from nearly white to a light blue, with little contrast between adjacent classes. Even where the data differ, the eye has difficulty distinguishing the second class from the third. A sequential palette with stronger contrast at the high end would have made the southern tracts visible even under equal intervals. Pale palettes are sometimes chosen to avoid alarming readers, and that motive may be reasonable, but the effect here was to remove the information the map existed to convey.

Choice 3: A Statewide Frame

The dashboard's map is embedded in a statewide view, with the county highlighted but small on the screen. Tracts in the southern cluster are a few millimeters across. A pattern that would be obvious at county scale becomes invisible at state scale. Monmonier (2018) describes scale and extent as among the most powerful, and least noticed, ways a map shapes what readers see.

The Three Together

No single change fully restores the pattern; classification matters most, but palette and extent determine whether a reader notices what classification reveals.

Table 1

How Each Choice Affected the Southern Cluster's Visibility (Composite Values)

Note. Composite values. Hatching marks tracts with fewer than ten diagnoses.

Was the Map Deliberately Misleading?

Nothing suggests intent. Equal intervals and pale palettes are defaults in some mapping software, and the statewide frame came from the dashboard's template. That is the more troubling lesson: a map can mislead through defaults, and a caption written after looking at the result can turn an artifact of design into a public conclusion. "No hot spots identified" describes the map, not the county.

Is It Fair to Redraw Someone Else's Map?

Redrawing is the most direct way to show a design's effect, and it is fair as long as the redrawn versions use the same data and the analysis says what was changed. All revisions here use the dashboard's own values. The aim is not to replace the dashboard's map with a more alarming one but to show readers that its conclusion depended on choices they could not see.

Why the Caption Matters Most

Of the dashboard's elements, the caption did the most harm. A reader who saw only a pale, uniform map might simply have moved on; the caption turned that impression into a statement of fact that board members could repeat. Captions on public health maps carry authority, and readers rarely check them against the data. A caption that described the map's method, such as "percentage of diagnoses found late, by tract, equal-interval classes," would have invited the question of why the classes were equal; a caption that announced a finding closed it. Monmonier (2018) notes that the words around a map often shape its reading more than the map itself, and this dashboard illustrates the point. The fix is simple and costs nothing: describe the map, not the conclusion, and let readers judge the pattern for themselves. A program that wants to report hot spots should do so with a formal cluster analysis, stated as such, rather than with a caption under a choropleth whose appearance depends on classification. A cluster method would at least state its assumptions and its threshold, which a reader could examine; a caption offers nothing to examine. Readers deserve the former, and a public dashboard that informs board decisions should hold itself to that standard. The revisions above would cost the program a few hours of staff time. That is a small price for a map the public can trust.

Recommendations for the Dashboard

Replace equal intervals with a classification suited to skewed data, and hatch or footnote tracts with few diagnoses so a single unstable value cannot set the scale. Use a sequential palette with clear contrast. Show the county at county scale, with the state view as a locator inset. And retire the caption; a dashboard map should let readers see the pattern and state its method, not announce a conclusion. Slocum et al. (2009) recommend that legends disclose the classification method, a simple step that would have prompted questions about this map.

References

Brewer, C. A., & Pickle, L. (2002). Evaluation of methods for classifying epidemiological data on choropleth maps in series. Annals of the Association of American Geographers, 92(4), 662-681. https://doi.org/10.1111/1467-8306.00310

Monmonier, M. (2018). How to lie with maps (3rd ed.). University of Chicago Press.

Slocum, T. A., McMaster, R. B., Kessler, F. C., & Howard, H. H. (2009). Thematic cartography and geovisualization (3rd ed.). Pearson Prentice Hall.