A framework for understanding and predicting the take up and use of social networking tools in a collaborative envionment

Abstract

Online collaborative environments, such as social networking environments, enable users to work together to create, modify, and share media collaboratively. However, as users can be autonomous in their actions the ability to create and form a shared understanding of the people, purpose, and process of the collaborative effort can be complex. This complexity is compounded by the natural implicit social and collaborative structure of these environments, a structure that can be modified by users dynamically and asynchronously. Some have tried to make this implicitness explicit through data mining, and allocation of user roles. However such methods can fail to adapt to the changing nature of an environment's structure relating to habits of users and their social connectedness. As a result, existing methods generally provide only a snapshot of the environment at a point in time. In addition, existing methods focus on whole user bases and the underlying social context of the environment. This makes them unsuitable for situations where the context of collaboration can change rapidly, for example the tools and widgets available for collaborative action and the users available for collaborative interactions. There is a pre-existing model for understanding the dynamic structure of these environments called the "Group Socialisation Model". This model has been used to understand how social group roles form and change over time as they go through a life cycle. This model also contains a concept of characteristic behaviours or descriptors of behaviour that an individual can use to make judgement about another individual and to create an understanding of a role or social norm that may or may not be explicit. Although studies have used components of this model to provide a means of role identification or role composition within online collaborative environments, they have not managed to provide a higher level method or framework that can replicate the entire life cycle continuously over time within these environments. Using the constructive research methodology this thesis presents a research construct in the form of a framework for replicating the social group role life cycle within online collaborative environments. The framework uses an artificial neural network with a unique capability of taking snapshots of its network structure. In conjunction with fuzzy logic inference, collaborative role signatures composed of characteristic behaviours can then be determined. In this work, three characteristic behaviours were identified from the literature for characterisation of stereotypical online behaviour to be used within a role signature: these were publisher, annotator, and lurker. The use of the framework was demonstrated on three case studies. Two of the case studies were custom built mobile applications specifically for this study, and one was the Walk 2.0 website from a National Health and Medical Research Council project. All three case studies allowed for collaborative actions where users could interact with each other to create an dynamic and diverse environment. For the use of these case studies, ethics was approved by the Western Sydney University Human Research Ethic Committee and consistent strategies for recruitment were carried out. The framework was thereby demonstrated to be capable of successfully determining role signatures composed of the above characteristic behaviours, for a range of contexts and individual users. Also, comparison of participant usage of case studies was carried out and it was established that the role signatures determined by the framework matched usage. In addition, the top contributors within the case studies were analysed to demonstrate the framework's capability of handling the dynamic and continual changing structure of an online collaborative environment. The major contribution of this thesis is a framework construct developed to propose and demonstrate a new framework approach to successfully automate and carry out the social group role model life cycle within online collaborative environments. This is a significant component of foundational work towards providing designers of online collaborative environments with the capacity of understanding the various implicit roles and their characteristic behaviours for individual users. Such a capability could enable more specific individual personalisation or resource allocation, which could in turn improve the suitability of environments developed for collaboration online.

Date of Award	2016
Original language	English