A Word on Mapping Bias

b8948a49-7b71-4a2a-b87b-da7d6b0c39b1

Level 1 (<525) | Level 2 (525-530) | Level 3 (530-535) | Level 4 (535-540) | Level 5 (<540)


So you come across this map on an article relating to an issue or policy you’re a bit skeptical of. You see the geographic correlation in the map—the number of units (in parentheses) is very low in the northern Mountain States and Dakotas. If you’re like most people, you gloss over the details of the map and cede over that the author might have some sort of a point. After all… you can see that the data sources the map was written by are credible—the author seems to have provided objective proof.

Sadly, this lulls a lot of us into falling prey to an argumentum ad verecundiam—a faulty appeal to authority.

Let’s say that the author is trying to use this map to prove that Policy A is superior to the newly implemented Policy B. He then points out that in states where Policy A is most closely followed by, the number of units (unanimously unfavorable occurrences) is visibly low compared to other states.

Problem 1

A map like this is incapable of showing population, which is a huge factor in this map. The reason the northern mountain states and Dakotas have less unfavorable occurrences isn’t because they’re adhering to Policy A, it’s because they have comparatively low populations. The map shows raw figures, not rates—indicative of really charged bias.

Problem 2

If you look at the units in parentheses, you will find that they do increase in regular intervals. However, a careful glance at the color gradient will bring to fruition the massive disparity between increments in color and in unit. While the units come at regular stages, the colors certainly do not. The first four levels follow each other in shade—one by one; however, the final level’s color jumps four shades into the darkest color on the color palette. This creates in the reader a feeling that there is a greater disparity in states than there actually is.

Problem 3

The variation on the map is really, really small. The full range of data between states is only 15 units. Given that the numbers in parentheses are just a number of occurrences and not a score of any sort, you can tell that there’s only a 2.8% disparity between the highest and lowest levels of states on the map. This sort of variation amplification is probably the most common kind of bias you may encounter online.


cropped-loneumbrellalogo21.png

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

w

Connecting to %s