Is Facebook Really Bad for You?

[As ever, this is on the BBC News website]

The examination period is always stressful, both for those sitting GCSEs, A levels and the International Baccalaureate and for their parents and siblings who get ‘second-hand stress’ without even a certificate to show for their efforts.

My friends and I used to revise together, hoping that it would create enough social pressure to keep us working through the evening, but being in the same room is clearly no longer required. My daughter, in the midst of IB exams, and my son, facing GCSEs next week, have email, instant messaging and of course Facebook and other social network sites to keep in touch with their school mates and share revision tips and exam guidance.

Some revising schoolchildren probably found their access to Facebook severely curtailed last month, however, after The Sun revealed that those who checked the site every day dropped a grade in their studies while heavy users were doing as little as an hour of school work a week.

The story was far from exclusive to The Sun, as a quick search of Google News reveals. It made dozens of papers and websites, including The Times, The Calgary Herald, and The Australian, which told its readers that “Facebook fixation harms student grades” and referred worried readers to a Sydney University-based group called “I want to sue Facebook if I fail university”.

Social networking scare stories are becoming increasingly popular, perhaps because the internet remains strange and mysterious despite its popularity while the long term impact of the network on our society is only just becoming apparent. Journalists, who probably have more to fear from the growth of social tools and conversational media than most, may also be keen to highlight the dangers of the new technologies.

So we see absurd stories like the Daily Mail’s recent claim that “using Facebook could raise your risk of cancer”, stories that entertain and frighten readers in equal measure by combining carefully selected psychological research with unfounded speculation to create a tale that has no basis in fact but aligns perfectly with widespread fears about new technologies.

This could also explain the love-hate relationship with Twitter in the press, where the service is a dull and tedious celebrity circus one day, and a cool way to stay in touch the next.

Facebook’s impact on academic grades seemed to be different, however, as it was backed up by some real science. The findings were based on a survey of 219 students at Ohio State University carried out by doctoral student Aryn Karpinski and Adam Duberstein of Ohio Dominican University and presented to a prestigious meeting of the American Education Research Association, which is as scientific as the media gets.
But of course things are never as straightforward as they seem, and the research which looked so conclusive in the pages of The Sun is actually far from definitive.

Karpinski’s presentation, “A Description of Facebook Use and Academic Performance Among Undergraduate and Graduate Students,” was not an invited peer-reviewed paper but a less formal poster session at the conference. The data showing a correlation between Facebook use and academic performance had not been published, and most of the news coverage seems to have been based on reading the abstract of the session without looking at the detail.

The press coverage prompted further investigation, and in an article for the online journal First Monday Josh Pasek, eian more, and Eszter Hargittai describe how they analysed data from other studies to see if Facebook did have the claimed effect on grades. They found no support for Karpinski’s findings, noting instead that “if anything, Facebook use is more common among individuals with higher grades.”

Karpinski then defended her work, noting that “my exploratory study and subsequent poster presentation were very basic. I merely planned to do this… to get some ideas and network with more experienced researchers in this area.”

She also looked at the surveys that seem to invalidate her conclusions, pointing out what she calls “serious methodological and statistical flaws”, questioning the coding methods used for the grade point average (GPA) data used, and noting that:

if GPA is an ordinal variable with four categories… then the ordinal nature of the dependent variable warrants the consideration of other regression analyses such as the cumulative odds model, the continuation ratio model, the ordered probit model, or stereotype model.

I did A level maths and have a degree in psychology, and I understand enough of this to appreciate what she is getting at, though I’d have to head to a textbook to see what the various models she describes actually involve. But this is the point where speculation turns into real science, and understanding why coding methods and the choice of analysis matter to the final results is vital for any accurate reporting of what is going on here.

Of course that understanding and analysis is completely lacking from the press coverage. None of the newspapers and websites that were so keen to exaggerate the original claims seem interested in following the real scientific debate, with the honourable exception of Carl Bialik in the Wall Street Journal who was careful to discuss the limitations of the original research and said right from the start that the area needed more study.

The press move on to another scare story, the impression that Facebook is bad for your studies remains, and the detailed research that will help us understand the emerging network society remains unread and unremarked upon. Perhaps we will have to wait for the semantic web and intelligent search, so that anyone calling up a dodgy article about the dangers of social networking is forced to review the latest academic research before they proceed.

Bill’s Links

The Sun story:
Daily Mail says Facebook could give you cancer:
Ben Goldacre takes on cancer scare on Newsnight
danah boyd on Karpinski:
Critique of the Karpinski research:
Karpinski addresses the criticism:
Carl Bialik in the Wall Street Journal: