Community Indicators for Your Community

Real, lasting community change is built around knowing where you are, where you want to be, and whether your efforts are making a difference. Indicators are a necessary ingredient for sustainable change. And the process of selecting community indicators -- who chooses, how they choose, what they choose -- is as important as the data you select.

This is an archive of thoughts I had about indicators and the community indicators movement. Some of the thinking is outdated, and many of the links may have broken over time.

Monday, April 16, 2007

How to Lie with a Graph

This morning, I received a copy of the Heritage Foundation's graph (PDF) "providing the facts" of tax cuts at work. If you've read Edward Tufte's work, then you can see what the problems are with how the data are displayed.

In The Visual Display of Quantitiative Information, we learn about the "Lie Factor", which he defines as:

Lie Factor = size of effect shown in graphic
-------_____---------size of effect in data

He then makes this point (p.57 of his book, if you're following along):

"If the Lie Factor is equal to one, then the graphic might be doing a reasonable job of accurately representing the underlying numbers. Lie Factors greater than 1.05 or less than .95 indicate substantial distortion, far beyond minor inaccuracies in plotting."

Let's look at the Heritage Foundation graph, then. It's on a PDF file, so the proportions are not skewed from the intended display.

The graph shows the number of jobs from January 2000 to January 2007. The January data points are provided, along with a drawing of a worker. The y-axis is scaled from 132 million to 146 million.

The low point on the graph is January 2002, with 135.7 million jobs. The high point is January 2007, with 146 million jobs. The percentage change is 7.59%.

The figures, however, tell a different story. The January 2002 worker is 3/4 inch tall. The January 2007 figure is 3 inches tall, a 300% difference. The "Lie Factor" (300/7.59) is 39.5, which is outside the range of 1.05-.95.

There's an important story to tell with the data on job creation. But distorting the data doesn't provide the facts or tell the story accurately.

1 comment:

  1. It's amazing how easy that is to do just by changing the starting point for the y-axis from 0 to 132. The change from 135.7 to 146 is only 7 percent, but they graph it essentially as a change from 3.7 to 14, visually representing a 300-percent change. "Lie" indeed seems like an accurate description.