The internet, depression and drinking a glass of water

A new study has made headlines around the world that claim that internet use is linked to depression despite better evidence from previous studies that there is no substantial link.

The study itself is a fairly straightforward online survey with the key finding that out of 1,319 people who completed the questionnaires, 18 were identified as ‘addicted’ by Kimberly Young’s Internet Addiction Questionnaire and these people were more likely to score highly on the BDI – a standard questionnaire to measure depression.

The study itself was well conducted although it is not a surprising finding because Young’s Internet Addiction Questionnaire (which you can read online here) asks lots of questions about emotional distress, so it’s hardly surprising that people who say they’re distressed on one questionnaire will say they’re distressed on the other.

I have criticised the concept of internet addiction on the basis that the whole concept doesn’t make sense, but research has also shown that these ‘diagnostic’ questionnaires are not particularly reliable, meaning they not a good guide even to what they claim they’re doing.

But perhaps the most important point, is that this study is just one in a long line of studies that have looked at whether internet use is linked to changes in mood.

Recently, a type of study called a meta-analysis was published that looked at all of these previous studies to see what the overall effect was – in essence, a mathematical aggregation of all the reported findings to get at the big picture.

This meta-analysis found that there was a statistically reliable link between internet use and depression, but one so small as to be insignificant. In fact, it found that internet was responsible for between 0.02% and 0.03% of total changes in mood (stats geeks: the variance was not reported directly but I calculated it from the r by the coefficient of determination).

In other words, internet use explains so little of a person’s depression that it’s irrelevant. It’s like knowing that hypothermia is a serious medical condition and that drinking a glass of water reliably lowers body temperature, but by such a small amount as to be medically unimportant.

Interestingly, I am quoted in some of the news stories about the study. Actually, I was contacted by a BBC journalist and some other stories have seemingly just nicked the quotes (often wrongly describing me as a psychiatrist).

What’s curious is that I sent the BBC journalist a link to the meta-analysis, even explained what it found and what a meta-analysis is, and included comments about why the study doesn’t change the general conclusion.

Instead of focusing on the existing evidence, I am quoted as being a naysayer. I have not been misquoted but the most important scientific point is omitted at the expense of presenting my words. This seems to be a common pattern where news stories often privilege opinion over data, when science privileges data over opinion.

In fact, the motto of the Royal Society, the world’s oldest scientific society, translates as “on the words on no-one”, but news stories often turn the hierarchy of evidence on its head, giving a skewed impression of the most fundamental way in which science works.

In this case, to suggest that science has established that internet use is strongly linked to depression when we know that it isn’t.

Link to PubMed entry for latest study.
Link to PubMed entry for meta-analysis.