At a glance, you may not notice any problems with the graph. Fortunately, Corpus Callosum exposed some of the major problems with the chart. I will summarize them here:
- The Y-Axis does not begin at zero. That happens far too often in graphs and is the least troubling mistake of the whole graph. You can counter this problem by labeling the Y-axis.
- The Y-Axis does not show job losses, but rather the total number unemployed at each data point. If your title suggests job losses per quarter, then your Y-Axis should relate to that.
- The X-Axis does not show quarters, but rather an uneven look at months. DEC '07 to SEPT '08 spans 9 months or 3 quarters. SEPT '08 to MARCH '09 cover 6 months or 2 quarters. Finally, MARCH '09 to JUNE'10 lasts for 15 months or 5 quarters.
- The graph shows a linear relationship. They accomplished this trick by poor axis labeling as opposed to accurate data analysis.
Once again, only a glance at the graph would imply that over the last quarter we have lost 15 million jobs. However, the Bureau of Labor Statistics actually features a much more accurate graph. While the job market still continues to prove a major problem to the economy (a point that doesn't even need to be exaggerated), creating a dishonest graph begs dishonest motives. Here is the real graph relating to the Number of Unemployed by Month.
You will notice a few things:
- The Y-Axis does not start at zero. At least they labelled the Y-Axis to show the starting point.
- The graph does not show a linear increase in unemployment from 2007 to 2010.
- Since mid-2009, unemployment has steadied.
No comments:
Post a Comment