A new search engine feature may allow political groups and individuals to alter the perceived popularity of candidates in an election, such as the recent race to fill Ted Kennedy's Senate seat, according to a paper by two academics at Wellesley College in Massachusetts.

In addition to its regular search results, Google recently began including up-to-the-minute "real-time web" traffic on the first results page, including breaking stories from news services and related messages from social networking websites like Twitter. This new function picks out unverified information solely based on how widely it is disseminated, and may not always be trustworthy, said Wellesley computer science experts Professor Pangiotis Taxis Metaxas and Eni Mustafaraj, a visiting scholar, in their paper, "From Obscurity to Prominence in Minutes: Political Speech and Real-Time Search" (pdf).

By meeting the loose criteria for this new feature, miscreant users are able to push an unchecked political agenda by crafting hundreds of Twitter messages that take advantage of unsuspecting users' likelihood to click on included links and forward the tweet to others, the authors said, raising the message's profile. This can widely spread misinformation about candidates or falsely enlarge apparent voter support.

Metaxas and Mustafaraj analyzed more than 185,000 messages posted by almost 40,000 Twitter users about candidates Scott Brown and Martha Coakley in the seven days surrounding the recent January 19 election for the Massachusetts Senate seat. Of these users, 205 posted more than 100 messages each, with some tweeting over 1,000 times.

Source: From Obscurity to Prominence in Minutes: Political Speech and Real-Time Search (pdf)

That sure is a lot of tweeting.

One out of every three of these messages, almost 60,000 tweets, is a repetition, or "retweet," of another message. Since most of the users read and share tweets among others of a similar political ideology, Metaxas and Mustafaraj suggest that the users' goal is not to emphasize a point, but rather to garner the criteria to be included in Google's new real-time web.

They wrote: "We believe that this fact shows awareness of the new role that real-time media plays, since it does not make sense to bombard your followers, with whom you greatly agree, with the same message."

The authors do not prove that Twitter users are specifically targeting real-time web search results, particularly since, as they point out, Google only announced this new feature a month before the Massachusetts election. However, they do delve in to some of the other repercussions of the widely disseminated tweets Google picks up. "Spikes in Google searches will attract media reports that attribute such spikes to predictive power," they wrote, indicating that perceived popularity among Twitter users broadcast on the real-time web will lead the media to incorrectly report candidates' actual popularity.

Metaxas and Mustafaraj further suggest that some political organizations may be targeting Twitter's free connectivity to broadcast spam to influence the electorate by increasing unwitting traffic to specific websites. For instance, they found a collection of nine similar usernames created within 13 minutes of one another all including "Coakley" and negative language. They believe these usernames automatically sent out almost 1,000 tweets to more than 500 other users in just over an hour and a half. The messages included text such as "AG Coakley thinks Catholics shouldn't be in the ER, take action now!" along with a links to a related website unsigned by its creator, also produced the same day.

By utilizing different usernames and various combinations of websites and twitter text, the spammers successfully dodged Twitter's spam detection and got other users to forward these messages.

"While we cannot know how many of these users either read or acted upon these tweets," Metaxas and Mustafaraj wrote, "the fact that a few minutes of work, using automated scripts and exploiting the open architecture of social networks such as Twitter, makes possible reaching a large audience for free...raises concerns about the deliberate exploitation of the medium."

Metaxas and Mustafaraj do not prove that Twitter users, malicious or not, are purposefully broadening the scope of their messages specifically to get on Google's real-time web. However, they certainly demonstrate how easily tweeters can manipulate the site's connectivity to propagate their own cause - a good perspective to keep in mind for the upcoming mid-term elections.