Micah Sifry has a great post at TechPresident on the use of Facebook “sentiment analysis” to explain political campagin trends. The gist of the post is that looking at whether negative or positive terms are associated with a candidate on Facebook is of little usefulness because it is impossible to detect irony. So if someone posts that they are “happy” Newt Gingrich is still in the race, it’s difficult to know if that person is saying it as a supporter or as someone who wants Mitt Romney to win. Without knowing other things about the nature of the Facebook user, it becomes impossible to know what “happy” means.

However, in context, it becomes much easier to predict what a poster means when they say they are “happy” a candidate is still in the race. The amount of content individuals post about themselves on social media sites like Facebook or through the profligate use of Google’s array of cloud applications is staggering. By taking all of that data, “data ninjas” can create startlingly accurate models that predict individual human behavior. It has heretofore been difficult to predict behavior in the social world because there is so much individual variability (the problem of inferential statistics as being “mean centric”). But the plethora of self reported data makes it increasingly easier for statisticians to create accurate models of individual behavior. If you are the unit of analysis, then accurately predicting your next move is simply a matter of having enough data. For all of our presumed spontineity, we humans rely a great deal on regularity and routine.

The implications of the product of what I call in an upcoming book the “fully specified self,” where marketers can with increased accuracy predict our behavior and provide us with opportunities to engage in that predictable behavior, are profound. Recent books by Eli Pariser and Jarod Lanier do a nice job detailing the problem with interacting with a medium (the Internet) whose central purpose is to give you back to yourself. I fear that for most of us, the natural response is to become more rigid in our core beliefs.

I wonder if the appropriate response to this is to intentionally break your routine (e.g. post something you disagree with or listen to music from a genre you’ve never heard of). In politics, this might mean consuming media from contrary views. For all the desire we might have to “feel right” about our belifs, we have an important contradictory impulse to live authentically. We want to be right, but we also want our beliefs to actually be right. The Internet does intersting things to this tension, giving us fodder for believing that what we think is right is actually right.