Dr. B's Blog

Ψ Bad research or poor journalism?

Welcome to the Wall of Shame, Yorkshire Evening Post. You are the inaugural post. You nailed it so well that I created this award just for you. In one article, you did everything right/wrong.

  • Sensational headline? Check.
  • Unsourced “research”? Check.
  • Picking on a highly stigmatized group? Check.
  • Cherry picking what’s likely a small paragraph in the discussion section and making it the focus of your article? Check.

It’s articles like this one that made me really start this blog. The title of the article reads “Internet surfers caught in a web of depression.” Oh my god! The carnage and brutality we have inflicted upon our fellow man!

And the opening sentence is even better

A “dark side” to the internet suggests a strong link between time spent surfing the web and depression

I’ve got several problems with this article but the first has to do with the lack of citations. How am I supposed to determine whether the research methodology was sound without a link to the original article? I suppose I can’t with this

Source: n/a

at the bottom of the page. I’d like to believe that research conducted on the internet can be valid (I did my dissertation by collecting data online, after all). In other words, I know how hard it is to get a valid sample on the web.

But, let’s assume that they’ve got a valid sample. The article above pulls out one particular juicy tidbit from the research

Her team identified a small group of 18 hard-core internet users who spent many hours online each day and were classed as “internet addicted”.

Their average depression score was more than five times higher than that of non-addicted users, and they were more likely to be moderately or severely depressed.

Wait. They found a (very small - 1.2%) sub-sample that was more depressed than the rest of their sample? Wow! What groundbreaking research! (That’s sarcasm in case the keyboard didn’t convey the message.) They had a reasonably large sample size and then they try to draw conclusions from 18 participants? That type of work wouldn’t have garnered a passing grade in my graduate school research design class and I wouldn’t even consider submitting a poster to a conference with a sample that size.

Here’s the thing: these findings may be legitimate. We know that spending too much time doing anything that’s not social can contribute to social isolation. (Please read that sentence again. See what I did there? I can do this “research” stuff too!) And that social isolation is correlated with depression. But to put this type of information in a mainstream media outlet? I don’t know if it’s necessarily bad research but I know that’s just irresponsible journalism.