Correlation does not imply Causation


It bears repeating once again that Correlation does not imply Causation. We are back again, with the dueling charts that attempt to show one political party in the United States is better than another political party based on government statistics plotted against who is in power. The first chart to the right is one that we talked about some time ago. Speaker of the House, Nancy Pelosi, would have us take a look at the job losses as report by the BLS, and correlate them with the President of the United States. From the bar chart to the right, one could mistakenly make the case that President Bush caused all the job losses, and President Obama has stopped the job losses.


However, if we look at the cumulative job losses, then one can make the case that things have continued to get worse under the Obama Administration. The chart to the right looks at the cumulative job losses under both administrations. The Y-axis to the right shows job losses in the thousands of jobs. Thus, the cumulative job loss is around 6.8 million jobs. Both the first chart, and the second chart use the same data from the same place. Both tell a completely opposite story. Neither is correct since there is no real link between who is President and job losses. It bears repeating once again that Correlation does not imply Causation.


Today, we see another one of these charts that attempts to paint one party as being better for jobs and the other party as being a jobs killer. Matthias Shapiro has created the third chart to the right, and he makes the same mistake that the other charts have made. He looks at whether one party in Congress, as opposed to the President, creates more jobs than the other. As British Prime Minister Benjamin Disraeli said “There are three kinds of lies: lies, damned lies, and statistics.” From the article:

The chart, below, is Shapiro’s graphical attempt to answer the question, “Does a Republican congress create more jobs?”

Shapiro gathered data from the US Bureau of Labor Statistics. Employment numbers are averaged by quarter and charted from 2003 to the present. (2010 Q1 is just January, 2010) Republicans took control of both houses of Congress in January 2003. Democrats took control of both houses of Congress in January 2007.

via Data Visualization: Fighting Fire with Fire | TechRepublican.com.

PG

This story written by Paul Adams

Paul Adams manages an award-winning team for a federal high performance computing facility where he has worked for 15 years. He loves getting his hands on the latest visualization and computer hardware, astronomy, aerospace engineering, working with the poor, and ringing cowbells.

Graphics , ,

  1. March 12th, 2010 at 09:46 | #1

    Thanks for the pointer! If you re-visit my blog, you’ll see a blog post debunking my own chart as well as an explanation of why I created a chart that I could then debunk.

    Like or Dislike: Thumb up 0 Thumb down 0

  2. Randall Hand
    March 12th, 2010 at 09:58 | #2

    @PoliticalMath Thanks for coming by to let us know :) I love your comments and agreen completely, Viz (like statistics) can present the data any way you want.. It’s very difficult to create truly ‘unbiased’ visualizations.

    Like or Dislike: Thumb up 1 Thumb down 0

Please consider using one of our verified authentication methods from the left when submitting comments.
* = Required item
VizWorld.com is a production of VizWorld, LLC © 2009