A Case Study In How Infographics Can Bend The Truth

These remarkable charts from Hyperakt are all "true," but each one presents a very different picture of reality.

We’ve made the point time and time again that charts and graphs, though they feel official and true, can lie. Rarely do you get to see that at work, but the good folks at Hyperakt have sent us a prime case study in infographic deception. The subject, of course, is politics—and in particular, the raging debate over whether the rich should be made to pay more taxes. "Using the same data, very different stories can be told depending on different agendas," says Deroy Peraza, one of the founders of Hyperakt.

A story from the Wall Street Journal's far-right op-ed page gets us started, with a chart showing how much taxable income is made by Americans ranging from the rich to poor:

Looking at that, the conclusion seems glaringly obvious: The rich don’t make so much money! In fact, as a group, they seem to make about as much as the poorest Americans! Why on earth would you tax these poor souls? As Jonathan Chait points out:

The chart most certainly does not demonstrate the Journal’s point. It instead relies upon an optical illusion. Democrats have been arguing that their tax increases should solely effect income over $250,000 a year. The Journal makes that pot of income appear small by diving it up into seven different lines. See, the $100,000-$200,000 line is tall, and all the other lines to the right of it are short. That tall line must be where the money is!

And look closer: The left side of the chart deals with people who make between $0 and $50K. So that’s a $50,000 band of income. Meanwhile, the right side of the chart deals with people who make between $500K and over $10 million. Which is a band of $9,500,000 plus. Those two don’t seem terribly equal do they? Why would you give over equal space representing each band? That’s not an honest approach.

When you look at the same data another way, you’d draw a totally different conclusion. Here’s what the chart looks like redrawn by the left-leaning Mother Jones.

Whoa whoa whoa! The rich make all the money don’t they? We should definitely be taxing those smug bastards. But step back again, and you realize this chart is misleading too. All of those making $200K and over are lumped into a single bar, while everyone else gets broken out into relatively tiny increments. Of course it would look like the rich make all the money. You could actually imagine doing this same chart in reverse, making it look like the poor make all the money, simply by lumping together everyone with less than $50K in income, and separating out as many groups of rich people as you please.

Surely the reality lies between these two extremes. And that’s exactly what Hyperakt has tried to capture, in a redrawn chart that shows how the IRS defines income. Here is the very same data grouped according to tax bracket:

At last, this seems like something close to an objective way of looking at the situation. And indeed, what it suggests is actually quite rational: Sure the rich don’t make all the money in the U.S. But they do make a very large portion of it. Even still, the chart does still fail to capture a couple more facts.

For one, basic economics tells us that a dollar made by someone making $30K a year means a lot more to that person than a dollar made by someone making $1 million a year. Ideally, you’d want a chart that would explore that feature by examining, for example, what exactly the rich and the poor are buying with their income.

And another point: The chart doesn’t show you how many people lie within each bracket—and if it did, you’d realize how very concentrated that right-most bar is. The fact is, the topmost one comprises less than 10% of the population. Which is to say that among the roughly 140 million American taxpayers, over 126 million fall within the two brackets at the left, while less than 14 million make up the bracket to the right.

"As designers it’s our responsibility to find honest ways of representing data so it isn’t used to mislead," Peraza points out. But likewise, we should all recognize that finding the truth of a situation is an iterative process. It takes time, and trial and error. Our first few attempts at capturing what’s true are almost never right.

Top image: S. Chiariello/Shutterstock

Add New Comment


  • Edward Nowak

    Why are sites on the left almost always only "leaning" and sites on the right "far"?

  • Dvanliere

    To EDN618 - Jai's answer, notwithstanding, it has to do with the perceptions and beliefs of those framing the questions...and I believe they, like Jai, believe what Jai said...things coming from the left are from people who believe the views coming from the right are 'far.'  On the other hand, while not in the print media as much, when you 'listen' (much of the venue available to the right) to the opposite views, any view from the left is 'far.'  We have not taken time to understand where others are coming from, why they believe what they believe, and what they hope to accomplish.  We simplify others to some common denominator of our choosing (the Sound Byte generation) - blogs are one or two phrases (not even sentences) per 'paragraph' because people believe otherwise no one will pay attention.  There is little thoughtful discourse and each side is pandering to emotion more than reason.  We talk about the politics of hate because that is what it is.  We lump people with extremists in the news everyone can identify with because it is easy...and effective because we're trying to wind the next election, not improve the system.

  • Jai

     Because, our political discourse in the US is shifted right of center.  There's a whole world of "far left" revolutionary Marxists who aren't really in he discussion.

  • Edward Nowak

    Thanks for the opinion, Jai. I completely disagree. Are you aware of any journalist surveys or media analysis (serious analysis, with peer-reviewed studies) which shows a bias to the right? I'd love to check any out that you could refer me to, because every media study of the topic which I've ever seen has shown a liberal bias. Two quick examples for you: journalist voting tendencies and political donations. They self report overwhelmingly as democrat voters, and political donation records show the same overwhelming democrat domination. They write the news, and intentionally and unintentionally incorporate their political biases, hence liberals "lean" and conservatives are "far."
    Your second point, if true, really frightens me, because I hear quite enough from the Marxists who ARE in the discussion.

  • Tim Anderson

    Nice article, thanks for sharing. 

    I hope the folks over at Mashable will read this as I think they are one of the worst offenders in spreading iffy infographics. It seem to be a 'cool' trend out on the intrawebs; however, I suspect most infographics are more about the graphics and less about the info.

  • Rick Yvanovich

    Love this post Cliff - thanks.
    Gives great weight to "lies, damn lies and statistics" or rather just how powerful info graphics are to communicate a message.

  • Adam Kaplan

    If the Hyperakt graph set the bar thickness proportional to the number of individuals in the bucket, you could potentially convey how many people generated said income.

  • Patrick Smith

    people that don't understand data classification shouldn't be making charts and infographics

  • Guerrilla Cheese Marketing

     I wish this was an infographic of why infographics can bend the truth..

  • PinPointPlanningLLC

    Cliff's last two sentences sum it up nicely. As a researcher I've worked alongside designers for years and very few get caught up in a "one size (data) fits all" method...in fact, most designers are reluctant to settle even after many iterations! The core contribution of good design is providing context to data - an under appreciated talent, I'd say.

  • Steve

    Infographics are just statistics and statistics have always been able to be 'massaged' to show a different picture.

  • Patrick Donnelly

    This has always been the ethical concern with data viz , no ?  Its a design obligation.    @pdonnelly01:twitter  

  • Steve

    Cliff, what does it matter what people earn, unless they received their money through violence or fraud? As long as transactions are voluntary and consensual (no threats of violence or fraudulent information), what claim does a 3rd party (the government) have on the earnings of free trades between people? I think the emphasis should be on cutting down the cost of government (who we only have 1 in 100,000,000 say in determining how they spend our money), not trying to extract more from the private consensual transactions. In free trade, both sides win because by doing the trade, they demonstrate that it is beneficial to not doing the trade. When government takes money, they don't have to prove they did any service for us, they just take what they want and try to get the majority to give them social validation for the theft. It really would become a beautiful world when we can pay for what we want, and never be forced by threat of imprisonment (slavery) to pay for wasteful wars, welfare or bailouts.

  • Mikal Christopher SCOTT

    In many, if not most cases, the information NOT displayed provides the direction that the creator wants the viewer to take in the informtion...
    as Robin pointed out, critical data that is missing gives a more 'accurate' display of the actual data vs  the 'dumbed down' version that is provided to the general public.

    I think term 'grain of salt' might work...

  • Patrick Donnelly

    Yes, there is a fine line between leaving off the noise vs leaving off the stuff you just dont want to acknowledge -    @pdonnelly01:twitter  

  • Rick

    You call the WSJ "far-right" but Mother Jones is only "left-leaning"? Way to leave your credibility on the floor. What outlet would be "center" - MSNBC? What would be "far-left" - the Daily Worker?