Initial stab at my thesis topic

Here's some material I posted as my preliminary thesis topic last night. I would absolutely love feedback on this -- does it make sense? Where do I need to provide explanation/explication? Does the whole thing seem doable as a project?

How do individuals assess the humanness of an organizational avatar in a social network site, and what factors determine the suitability of these avatars within the context of their chosen site?

A couple years ago, college admissions offices started having freshmen blog about their experiences for the benefit of prospective students. Having worked in university communications, I feared that the resulting blogs would be tightly controlled by public relations offices and generally stripped of their humanity. Although I was pleasantly surprised by the freedom afforded these admissions bloggers, I started to think about how individuals visiting these sites constructed a mental impression of the personalities represented there. I found a paper describing an encounter between an individual and a chat bot particularly interesting—could the act of determining humanness become the crux of marketing, given that humanness is no longer assured when dealing with online interaction?

Fast forward to the present day, when organizations (really an automated program, an individual or a team of individuals) are establishing presence on Twitter, a social network site. Organizations do not have the same level of control over their presence as they would on their own website; they must work within the milieu of the social network site to convey their reputation. This can be challenging because Twitter does not distinguish between accounts representing an individual and accounts representing an organization.

How does an individual determine—via a sort of Turing test—whether a human, several humans or a computer program is behind an account on Twitter? If an account is judged to be human (or not), what causes an individual to choose to connect to that account, and what factors (such as social utility) influence whether or not that connection is maintained over time?

Recently, signaling theory (from biology and game theory) has been used to evaluate reputation and trust in virtual communities. A large amount of messages sent on Twitter are publicly available, and it is possible to reconstruct (or at least quantify) exchanges between two accounts. I plan to use signaling theory to identify the different cues provided by organizations on social network sites and how they are interpreted by individuals connecting with those organizations.