New word from the Pew Internet and American Life project is that the internet isn't driving a wedge between family members after all. Quite the opposite actually. It turns out mobile devices, digital media, and the internet are giving families new ways to keep in touch, new activities to share. Pew does really nice surveys, and this report has been written by some smart people. (Here's an overview of new findings from CNet.)

I hope, then, that this will really put an end to anyone talking about the HomeNet study. Back in 1998, a group of researchers from CMU published a study (called HomeNet) that claimed, among other things, that the Internet increased social isolation. In fact, they claimed all sorts of negative social psychological influences. Of course, the media took the story up with gusto, writing headlines like 'Sad, Lonely World Discovered in Cyberspace' (from the New York Times). Oh no, the world will come crashing down.

But wait. Just a few years later, the researchers re-did the study with a new sample, and found that nearly all the negative influences had disappeared. Hallelujah, we're saved! Rather than refuting their original results, however, the authors explained the reversal by noting that the internet had changed, and network effects combined with broader diffusion had eliminated the isolating effect of internet use. Hmmm. That might be part of the story.

But if you think that explanation seems too simple to be complete, I'd agree with you. Last year some colleagues and I re-examined the original HomeNet data, which the study's authors kindly make freely available. We applied a technique called Matched Sampling in order to look at the influences of internet. Some day I'd really like to get around to publishing the results of our work here. To make a long story short, when you compare apples to apples – like people to like people – you find almost no negative influences of internet use in the original HomeNet data.

Now, I would never state this as a definitive result. Our method has flaws too. But it's a good one, and just as carefully done as the CMU folks'. When the results differ in that situation, I start to worry that the results are model-dependent, and not reflecting a true inference. So, it's not that the internet changed, though it most certainly did, it's that the original inference might have been wrong. And that study was done by some very smart, careful researchers who used excellent statistical methods.

Why do I tell this long, geeky story now? Because I think refuting the HomeNet results is important for understanding the history of the internet. We should stop telling a story (if anyone is actually telling it), in which the internet isolated people at first, and then tapped them in. The internet is a case study in the diffusion of a new communication technology, and we should get the story straight. These new Pew data show what I believe has been true all along: people who are more social offline are more social online. The internet, mobile devices, etc. give people more ways to communicate. And rather than diffusing that communication over many new people, most people are using email, SMS, MySpace, Facebook, Twitter, etc. to communicate with people they already know: their family and friends.