In a previous diary, I referenced that wonderful 1954 book How to Lie With Statistics by Darrell Huff. A short fun read, and every technique Huff described 60 years ago is in regular use today.
One of the most common ways to exaggerate or underplay an effect on a graph is by slicing off just a portion of the number scale to give you the effect you want. In the December example, from Gallup, shortening the x-axis scale was used to make the fall in Obama's popularity look more dramatic than it was.
In the Fox graphic on ACA enrollments, the same technique is used to make it look like ACA enrollments are wa-a-a-a-y below original estimates. Of course if the scale started from zero rather than around 5.5 million, the bars would be much closer together, showing the goal as being almost reached.
http://talkingpointsmemo.com/...
Check out Huff's book for a great read. You may find it in your local library, or it's available on Amazon.