Misleading Statistics: How Visual Data Can Do Bad

Visual content like infographics are more popular than ever. They are a fantastic, shareable way to provide useful content to relevant audiences.

But as with the rise of any popular medium, for every exceptional graphic there are a number of poorly executed ones – infographics that are heavy on facts and data and woefully short on context, utility and relevance.

While there are an increasing number of tools to help create great-looking visual content, it’s just as important to focus on the information you’re conveying. Don’t overlook the basics of clear and accurate data presentation, like in the examples below.

Bad-Data-Size-Matters-In-Surveys

To paraphrase Twain, there are three kinds of lies: lies, damned lies, and what clutters up a lot of infographics. While we’ve seen “industry trends” proclaimed based on a survey of a few dozen individuals, be sure your data source uses a sample size more representational for confident accuracy.

Bad-Data-Correlation-or-Causation

And by the way, did you know that a lack of pirates is causing global warming? Comparing concurrent trends and events is worth a look, but be careful about implying that one causes the other. Your site traffic may have doubled after you started blogging, but dig around before suggesting that the latter was responsible the former. (It probably was — blogging rocks! But do the research.)

Meaningful-or-Just-True-Statistics

Statistics are a funny thing — it’s entirely possible for them to be technically true and yet utterly unhelpful. Be sure to find the data that’s sensible. Comparing the number of police officers in California to the number in Rhode Island as an indicator of relative safety is misleading, given the states’ comparative sizes. Comparing police per capita would be more relevant information.

Skyrocketing-Tax-Rate-Increase

It’s very easy for visual content to go from unhelpful to misleading. A simple bar graph like the above can suggest something markedly different from the underlying data it represents.

What examples of visual data gone bad have you seen?