I like graphs. And, obviously, I write about politics for a living, so when an ad backing Connecticut Gov. Dan Malloy (D) popped up during the (lamentable, horrible) defeat of the (noble, honorable) Kansas City Royals in the World Series on Wednesday night, I couldn't help but notice the graph that was prominently featured, trumpeting the sudden drop in the state's murder rate.

That ad, from the Michael Bloomberg-funded gun control group Independence USA isn't online, but we got a still of the graph. It's introduced as part of the group's efforts to support Malloy's strong advocacy of new gun control measures following the shootings at Sandy Hook Elementary in Newtown.


This made me wonder: How accurate is this? There's no labeling on the vertical axis, so it's hard to know what the line represents. Or even if it represents anything. And so I decided to Fact Check Some Graphs™, starting with this one, as a "service" to a math-hungry public.

So what about murders in Connecticut? That's easy enough to find, thanks to the FBI's online database and news reports from Connecticut. Presenting: the Bloomberg graph, in context.

The number of homicides in 2013 was down from previous years. There's a lot of intentionality in how the graph is displayed; for example, it only shows a few years so that the recent drop seems all the more stark. (Had they gone back to 1990, as we did, it would both diminish the effect of the reduction and prompt the question: Why did the rate drop before, even without Malloy's new gun control measures in place?)

Moving on, we find Allen Weh (R), who is challenging incumbent Sen. Tom Udall (D) in New Mexico. Weh ran an ad slamming Udall for being a millionaire even while incomes in New Mexico dropped and the poverty rate rose. And it used graphs.

This is the one about incomes.


Notice the lack of a vertical axis again, and the lengthy timeframe. Udall wasn't in office in 2001, so the drop in incomes (which is apparently what this is) from 2001 to 2005 isn't explained. (Also notice, if you click to view the larger version of the chart, the use of some random email as an overlay. "We need to complete some important maintenance work ... firewall which requires the site to be offline during that time." And at upper right, a stamp.)

Here's the actual data on incomes in New Mexico.

Weh's ad appears to use household income data; the line pattern matches roughly. Personal incomes, though, have grown since the official end of the recession in 2009. The ad also makes the point that New Mexico trails the rest of the country on economic figures, which is true -- but we're comparing graph to graph. Weh's graph also floats above an invisible zero line, making the rise and fall seem more dramatic than it actually is.

Speaking of misleading, here's his graph on poverty.



That last bar, the 80,00 one, sort of climbs and climbs, finally peaking at about 424,000 before it cuts away.

424,000 what, though? Well, people living in poverty. The actual chart:


The actual number of people living in poverty in New Mexico in 2013 was about 448,000, up from 426,000 the year before. But 2012 was down from 2011. (For years prior to 2011, we applied the poverty rate to the population.) This bears no resemblance to the sudden spike shown in Weh's growing graph; it's not even clear what those numbers are supposed to represent.

Then there's the sort of generic trombone-bwah-bwah-bwaahhhh downward slide graphs used in ad like this one from Colorado Rep. Cory Gardner (R) in his race against another Udall, Sen. Mark of Colorado.


Udall's poll numbers are dropping, suggests the text and that overweight red line stumbling down the graph paper. And Udall's numbers have dropped, putting him behind Gardner in the final days of the campaign. But the real numbers don't look much like the Gardner ad. We flagged September 18, the date cited by Gardner, for reference.


So why did Gardner (or, more accurately, the media consultants that made his ad) use the graph? Because graphs convey seriousness and rigor. It's why members of Congress love to schlep giant charts onto the floor of the House. Don't believe me? they say -- well, here is the data to prove it. Who can argue with a graph?

Me. I love graphs too much to let them be abused. So I will also point out, Mr. Gardner, that the graph at right below demonstrates 99 percent, not the sketchy one in your ad. These things matter to voters.*



* Ha ha no they don't. Good night, everybody!