The conventional wisdom these days is that our attention spans are getting shorter and shorter. We are always looking down at our phones, not looking at the world around us. Conversations don’t occur because we need to check email or Twitter or post a picture of our meal on Facebook.

So how bad is our attention? Well, according to what appears to be the most commonly cited statistic, we have attention spans of about 8 seconds, slower than the average goldfish (I won’t even guess how one estimates the attention span of a goldfish). Problem is, that number is made up. I have found no credible evidence that this statistic is based on anything other than people believing in something they saw online.

Where does it come from? Turns out, most people cite a Spring 2015 report from Microsoft Canada Advertising (yes, advertising) entitled, “How does digital affect Canadian attention spans?” In it, the (unnamed) authors surveyed 2,000 Canadian respondents about their use of mobile technology, websites, and online games.

Interestingly, attention span length was not even tested as part of the study! Instead, the statistic is included in some kind of balloon graphic on page 6 of the report that references the website Statistics Brain.

Screen Shot 2016-01-29 at 1.35.11 PM

So let’s go to Statistics Brain. Here, we find the three important (and offending) statistics.

Screen Shot 2016-01-29 at 1.07.10 PM

The bottom of the page holds two sets of sources:

  • The first is a paper by Harald Weinreich and colleagues from 2008 that doesn’t actually test attention spans, but instead conducts a “web usage study with twenty-five participants” that examines online browsing behavior (this post contains a good review of the article). It’s also worth noting that the paper was published in 2008, so the 2013 statistic can come from here.
  • The second, listed under “Statistic Verification” header identifies the “National Center for Biotechnology Information, U.S. National Library of Medicine, The Associated Press.” First off, always be wary when a news agency is listed as the original source. But the NCBI and NLM seem like reasonable possibilities, so back in May I emailed NLM to confirm, and they responded with a thorough email that concluded with the following statement: “We have had similar requests on this same issue saying it came from various sources such as NCBI and cannot verify it.”

Screen Shot 2016-01-29 at 1.41.20 PM

An email to Statistics Brain asking for further clarification about these numbers went unanswered.

You’ve seen this attention span number used over and over again: Time says, “You Now Have a Shorter Attention Span Than a Goldfish”; USA Today says, “Our attention span now worse than goldfish’s”; and the National Post (Canada) says that, “Canadians now have shorter attention span than goldfish thanks to portable devices.” Hubspot (my favorite source of just using whatever number they happen to find on the internet), also used it in an infographic on marketing to the attention-shortened public. Most recently—and I mean last week!—Timothy Egan from the New York Times cited it in an article about his own lacking attention.

As you might expect (and where I have keen interest), these comparisons also show up in blog posts about presentations and data visualization. The Mobile Presenter included it in a post last year. Speech coach and presentation author Michelle Mazur used it in a blog post on audience attention, where she cited Sally Hogshead’s book, How the World Sees You. Hogshead relied on this BBC News article, which, as you might expect, was based on the Microsoft report, which is based on the Statistics Brain page, which is, well, you get the point.

And Ethos3, usually one of the better presentation blogs because they cite actual research, said the same thing in “How-To Conquer Short Attention Spans.” In an interesting twist, they inserted the link over the title of the Weinreich article, but actually link to the Statistics Brain site.

And it shows up in Tweet after Tweet after Tweet…..

Screen Shot 2016-01-29 at 1.59.43 PM

I’ve written about other bad statistics in the past (here and here), and this one is no exception. It’s not correct! Look, it’s not that I don’t believe we have shorter attention spans in the past because I’m sure we do, it’s just that I don’t have a reliable number to put on it. And, by the way, I like John Medina’s (from Brain Rules fame) advice to break up instruction/presentation into 10-minute chunks, which is based on actual research on instruction.

I find it funny how these all circle back to the same root source, with no one wanting to take a closer look at where this number actually came from. So, as I did when I wrote about the 60,000 Fallacy, when it comes to the 8 second attention span, please Internet, stop using it.