Playing with Political Data

February 19th, 2010 by rubenr

It’s pretty widely understood that you can arrange and manipulate data to make it show whatever you want, and you can do this to varying degrees.  This graph released by Organizing for America got me thinking:

The Rate of Changing Unemployment Rates (Obama v. Bush)

It’s a very effective communicator of the changing rate of unemployment, though it could be considered slightly misleading since throughout this entire period we’ve still seen a net loss of jobs.  That said, the changing rates are worth looking at, it means we’re likely going to start seeing a shift towards adding jobs rather than losing them sometime in the near future.

TargetPoint, a GOP consulting company, decided they’d retort this graph with some graphs of their own.  While I think it’s a valid criticism of the Obama data to point out that we’re still losing jobs (just at a much slower rate), TargetPoint created some graphs that are really distorting:

It’s powerful, but extremely disingenuous.  First off, the rate for Bush is the average for his entire presidency over 8 years while Obama has only one-year’s worth of data.  The high unemployment during Bush’s last year in office is thus weighed down by his other 7 years.  But that’s not the most disturbing part of this graph.  You wouldn’t know from looking at it quickly, but the x-axis is set at 2.5% rather than zero.  While its true that unemployment has never been 0%, shifting the axis this way without proportionally changing the scale leads to quite a deceiving result: it looks like the red bar is less than half the size of the blue bar, but the last time I checked 5.3% is more than half of 9.3%.

Moral of the story: beware of data visualizations.  Also, people should actively call out shenanigans like this when they see them.


2 Responses to “Playing with Political Data”

  1. Matt says:

    As you state “It’s pretty widely understood that you can arrange and manipulate data to make it show whatever you want, and you can do this to varying degrees. ” This is Very True. But as the graph shows unemployment going away in Obama’s term, that is simply because the people that are being removed from the graph aren’t finding jobs, but are no longer eligible for Unemployment, and stop filing with unemployment agencies. (Why spend time of my day filling out paperwork every week just to be counted on a chart). Rethink your criticisms.

  2. rubenr says:

    Thanks for the comment! While true that those not receiving unemployment and/or not searching for jobs wouldn’t be included in the data, it’s important to note that unemployment benefits were extended last year (Up to a year and a half in some states like California), so I doubt it alone could explain the trends we’re seeing.

Leave a Reply