Researching social media usage: right research, wrong conclusions?

Just a short post, friends, to rant a bit about a couple social media research posts I’ve just come across. As invaluable good research is on the uses and implications of social media, I’m sometimes bothered by the conclusions drawn from research. I speak not as a researcher myself, and must express my gratitude to those who get their hands dirty with data collection, organization, and processing. But we all know that research is frequently conducted in order to test a hypothesis — and that consequently, data lends itself to proving the hypothesis to be correct. And there is the fact that some of this research is reported with a flair for the headline, and so the blogs on which we discover internet research may often distort or even falsely report research findings for the sake of a good lead.

I have two complaints. The first deals with conclusions drawn, I think falsely, from research conducted around online communication practices and proximity. The research used Facebook and email habits of users, and concluded that the internet is not a “global village” after all, but that users in fact communicate with people they live close to. Well the research is interesting, but clearly Facebook is not the only way in which people communicate online. And Facebook is a social network for friends. Presumably, if one included the many ways which we communicate with people we don’t know: twitter, blog commenting, groups and niche networks, then geographical proximity would not look like the cause of communication. it’s not the research so much as the conclusion that bothers me. and not even that the conclusion — we communicate most with people close by — is a bad one (it seems to me an obvious one).

It’s the theoretical misstep that bothers me. And it applies to other and similar research efforts. Aggregate user activity, captured in data, are a problematic foundation from which to make claims (such as observations) about user motives and intentions. Even more problematic, in my opinion, is the use of research like this to explain these motives. Either the researcher, or the reporter (bloggers included), will often draw conclusions that are neither supported by data nor expressed by the data. I’m not a scientist or statistician, but it seems clear to me that a finding such as “we use the internet to communicate with people close by” neither refutes the internet’s ability to collapse distance; to link disparate and unrelated people, pages, and communication forums; to aggregate commentary around blog posts; to create followings around personalities (twitter) in ways that can subvert mass media’s control over image and messaging; nor proves that the global village was a utopian idea, but an idea only.

To make those kinds of claims, one would have to study not just our personal communication habits, but our habits around tweeting, “publishing” and posting, participating in groups, playing social games, subscribing to news, and much more. One would have to know how internet-based discourse networks grow and function. One would have to make cultural claims about access and exposure, oh, and participation, in media events, stories, videos, and other kinds of socialized news.

The village, local or global, is made up of far more than personal conversations. Villages have news, gossip, rumors, and secrets. So it frustrates me when research is used to collapse a concept, when research is used to support conclusions it couldn’t possibly contain (because they are outside the data), and when simple stories are created as a means to explain or tell research findings. I’m sure researchers themselves are troubled when this occurs.

The article that prompted this post:

E-mail Traffic Data Casts Doubt on Global-Village Theory
mentioned at Modern Metrix

If you think e-mail is making geographical distance less important, think again. A new analysis indicates that the opposite may be true.
….
Their conclusion is that far from reducing the importance of geographical location, electronic communication appears to have increased it, probably because people swap more messages with those they have personal interaction with.
….
A lot of thinking about economics and numerous business plans are based on the idea that society has become a “small world.” There may need to be some hurried rethinking if that premise turns out to be wrong.

And the second part of this rant is sparked by an over-simplified categorization of social media users found at AndersonAnalytics. I am glad that there are others interested in social analytics, and from a behavioral and psychological angle. This has been my bailiwick for a while now. I don’t have research to support my “insights” into user psychologies. Anderson floats user types based on a simple form on which users self-describe themselves in ways that unsurprisingly match the characteristics used to differentiate users.

Others have taken this kind of approach, and while I have my doubts about the reliability of self-described behaviors (we don’t always know the whys and wherefores of our actions), it’s again not the research approach that bothers me. It’s the idea that user types may be identified that explain user behavior and experience. This is suggested by the labels: “fun-seeker” (a form question supporting this must be: “Express my creativity”, which is an awfully strange way to self-describe “have fun,” not to mention that being creative and expressing oneself are different things altogether, that being creative doesn’t require an audience, and boring people express themselves, too).

It bothers me when generalizations are made on the basis of data inadequate to the generalizations concluded (our first case, above) or on the basis of data poorly collected (our second case, below). Granted, research in this field is often cited to tell stories that support the cases made by social media professionals. But for those who might take such conclusions as we see here to heart, I want to simply say that online user experiences, and real phenomena of communication, of relationships, of community, etc, are far more complicated than suggested here. If you are in the business of using social media, demand better research and reporting.

From: Anderson Analytics’ Seven Social Network Segments

“To understand the SNS market more broadly, Anderson Analytics has created and profiled seven segments of online Americans. Three SNS Non-User segments and four SNS User segments. The yellow circles mapped below represent the four SNS User segments. The blue circles represent the three SNS Non-User segments.”

This entry was posted in Design, SxD Theory and tagged , , . Bookmark the permalink.