AbstractDigital devices and services have become ubiquitous and persuasive. Hence, it is critically important to understand these digital interactions (or digital traces) as daily life involves rapid transitions between various offline and online interactions.
In this thesis, I explored how individuals change and adapt their behavior across a variety of digital systems through the lens of social role and identity theory perspectives computationally. Individuals naturally adapt their behavior across contexts. This social flexibility is integral to human interaction. For instance, we would not expect someone to behave identically at work and home. Here, I extend this concept to the digital world, where behavior and identity is examined across a variety of digital devices and systems utilizing new methods and approaches.
I first made the case for using new methods to better understand digital behaviors by demonstrating the discrepancy between objective technology use and self-reported behavior, which provides an example of an alternative method for psychologists. Next, I considered the dynamic nature of social roles online, for the first time, by utilizing supervised and unsupervised machine learning techniques. The results show how roles relate to leadership online, and which pathways users took to become leaders. I then analyzed user linguistic flexibility via computational methods (linguistic style matching (LSM)) across different contexts. This revealed that users typically diverged their linguistic style from the community. This happened to a lesser extreme within strictly moderated contexts. Finally, I undertook a qualitative approach to user experiences across online systems, which further showed how users negotiate their self-presentation and identity online.
Overall, this thesis confirms that new, data intensive, objective measurements can sit alongside traditional approaches. If these are adopted more widely, social psychology will move further beyond the lab when it comes to understanding how people live across digital and real-world environments.
|Date of Award||20 Nov 2019|
|Supervisor||Adam Joinson (Supervisor)|
- computational social science
- behavioral analytics
- human-computer interaction
- behavioral science
- online communities
- online behavior