LogicalDUAL-USE

Misleading Map

What it is

Using the geography of a map — area, projection, color breaks, raw counts by region — to imply a distribution of people, votes or risk that the land does not carry, so that empty space reads as a majority and dense space disappears.

How it works

A colored map colors land, and land is not people. The 2016 US county map is the standard case: Donald Trump won roughly 2,600 counties to Hillary Clinton's roughly 500, so the map is overwhelmingly red for an election he lost by about two points in the popular vote and won in the Electoral College. Worked: Alaska has about 570,000 square miles of land and 730,000 residents; New Jersey has about 7,400 square miles and 9.3 million. On a land map Alaska is some 77 times larger; on a people map New Jersey is about 13 times larger; each state gets one color. Monmonier catalogued the other levers: the choice of class breaks, which can make the same data show a sharp regional divide or none; raw counts by region, which make populous places look worse at everything; projections such as Mercator, which inflate high latitudes. The persuasion move is to let the eye's equation of area with quantity do the work. Honest alternatives exist — cartograms that size regions by population, dot-density maps, per-capita choropleths with stated breaks — and a presenter who chooses the land map for a people question, and describes it as showing the people, has chosen the impression.

Real-world examples

  • •In May 2017 President Trump displayed a framed county-level map of the 2016 result in the White House; the “land doesn't vote” rejoinder, and cartograms sizing counties by population, circulated in response and showed a country divided nearly in half.
  • •In November 2020 the Brookings Institution reported that counties won by Joe Biden accounted for about 70 percent of US economic output, and the figure was circulated by his supporters as a map-based claim of national weight; it weights places by dollars rather than by people or by land, and no more describes voters than the county map does.
  • •County maps of raw COVID-19 case counts in 2020 showed populous metropolitan counties as the darkest areas even in periods when rural per-capita rates were higher; the same data mapped per capita reversed the picture.
  • •Monmonier's How to Lie with Maps demonstrates that the same variable, cut into classes at different break points, can be colored to show a north-south divide, an east-west divide or a uniform country, with every version numerically correct.
  • •The Mercator projection shows Greenland at roughly the size of Africa, which is about fourteen times larger; maps of resources, conflicts or emissions drawn on it give the high latitudes a visual weight they do not have.

Ethical guidelines

Where the line is

A choropleth or county map is legitimate when it is labeled as showing area-based results, its projection and breaks are stated, and it is not offered as evidence about the number of people on each side; it becomes manipulation when a land-area map is presented as a picture of the population, when breaks or raw counts are chosen to produce a regional story, or when the population-weighted version is known and withheld.

  • ●For any question about people, map people: use per-capita values, a cartogram or dot density, and say which you used.
  • ●State the projection and the class breaks, and check that a different reasonable set of breaks does not reverse the story.
  • ●Do not describe a land-area map as showing where “the country” or “the voters” are.
  • ●Show the raw counts and the rates when both matter, in separate panels, rather than choosing the one that makes the point.

How to defend against it

  • ►Ask whether the map colors land or people, and whether the claim being made is about land or people; if they differ, the map cannot support the claim.
  • ►Ask whether the values are raw counts or rates, and what the class breaks are; then ask what the map would look like with per-capita values and evenly spaced breaks.
  • ►Look for the same data as a cartogram or a dot map; they are widely published for elections and epidemics, and the contrast is the lesson.
  • ►Treat any map-based claim about a majority — of votes, of output, of anything — as a claim about a denominator, and ask for the number of people on each side.

References

  1. Monmonier, M. (1996). How to Lie with Maps (2nd ed.). University of Chicago Press
    Class breaks, projections, choropleth design and the ways a numerically correct map can mislead.
  2. Cairo, A. (2019). How Charts Lie: Getting Smarter about Visual Information. W. W. Norton
    The 2016 county map, the land-versus-people problem and cartogram alternatives.
  3. Gastner, M. T., & Newman, M. E. J. (2004). Diffusion-based method for producing density-equalizing maps. Proceedings of the National Academy of Sciences, 101(20), 7499-7504
    The method behind the population cartograms used to redraw election maps by people rather than area.
Last reviewed
Suggest a correction

Detect Misleading Map in any text

Paste any message, email, or article into our free Manipulation Detector to see if Misleading Map or other techniques are being used on you.

Try the Detector FreeCreate Free Account