As Americans abandon landline telephones for cell phones, traditional political polling grows less effective: A Pew Research Center report recently found that polls using only landlines to gather data can overestimate support for Republican candidates by 4 to 6 percentage points. Meanwhile, social media sites such as Twitter continue to gain users. In response to this shift, political campaigns are exploring "sentiment analysis"-a way to systematically eavesdrop on the public conversation about a candidate by gathering data from blogs, Twitter, and other social networking sites, then using a program to analyze how candidates' messages and activities affect potential voters. For instance, a sentiment analysis program might pick up a public Twitter update that mentions a politician's name, then scan it to determine if the opinion expressed is positive or negative.
Businesses already use sentiment analysis to keep tabs on their brands, but political campaigns have been slower to adopt the practice. One problem is that these programs cannot easily sense sarcasm, which is rampant on the internet. Yet a British company called Linguamatics used sentiment analysis to predict within a single point that the Conservative Party would win the last election, and some speculate that the practice will be common by the 2012 elections.
Complicated home entertainment devices mean complicated remote controls. But now, the companies that make such products are thinking about eliminating remote controls altogether. Mitsubishi, for instance, decided against creating a touch-screen remote control, since it would raise the price of its television by several hundred dollars. Instead, the company-and others like it-is developing apps for smartphones and tablets (like the iPad) that can do the work of remote controls. But there are drawbacks: For instance, remote control apps tend to drain device batteries. And what happens when several family members with smartphones and competing tastes in television can all control the television channel?
What can Facebook tell us about how we make friends? A lot, apparently. A recent study published in the American Journal of Sociology tracked a class of 1,640 new university students to see how they made friends. Rather than tracking subjects' Facebook "friends"-since this relationship can be impersonal-the study looked at people who were tagged together in photos. What surprised researchers most was that the students were not attracted primarily to one another by race, as sociologists have long assumed. Rather, the study suggested that students are more likely to become friends with those they see often-roommates, or people who have the same major. Additionally, the students were two and a half times more likely to befriend people from their home state.
The researchers noted that Facebook may represent a highly effective source of information for future projects because of the amount of information users tend to disclose. Yet while the researchers gained approval from the university and Facebook to perform their study, students did not know they were being tracked, sparking privacy concerns that could plague future research.