Search

This study investigated online public political conversation in the UK. Drawing on theories of deliberative democracy, it emphasised the importance of inter-ideological discussion between citizens in the formation of informed opinion and preferences, focussing on the potential of the internet to facilitate this through large scale, ideologically diverse conversations. A multivariate analysis investigated the roles of interface design, institutional linkage and participant community dynamics in the formation of online political conversation. The investigation of conversation from across the internet required a very large scale approach, situating the study within the big data paradigm. However, it also required deeper understanding of human communication, gained through more qualitative analysis. Therefore the study utilised a novel, iterative, quali-quanti approach featuring initial, large scale quantitative analysis – involving bespoke software to automate the collection and analysis of conversation data – that was used to direct further iterations of increasingly smaller scale and qualitative analysis. Reflections on the successful application of the methodology are significant in themselves, but the study also generated novel observations of online public political conversation. The findings illustrated participatory spaces as unique online niches, each with specific communities and goals, and described how participant agency allows citizens to contribute according to various democratic models. For example, an action-oriented approach existed in policy related spaces, in which participants sought only to express a preference, rather than engage in discussion. In more discursive spaces, non-political social bonds between participants were seen to be particularly important in the facilitation of civil, productive, inter-ideological debate and certain participatory roles were important in facilitating these bonds. The design of spaces exerted a significant, but not determining effect on conversation, being used to present conversation in particular ways. However, certain features, notably active facilitation, helped to shape conversation through enabling some of the important community roles to be performed.

Bacterial flagellar filaments are long cell surface appendages that generate propulsion for movement. They also play key roles in surface attachment and host-bacterial interactions. The filament is made from a single protein species flagellin. Approximately 45 % of annotated flagellar systems possess multiple flagellin genes. We have investigated the ability of Caulobacter crescentus to build a flagellar filament using six flagellins: FljJ, FljK, FljL, FljM, FljN and FljO. Our analysis showed that this flagellar system exhibits extensive structural redundancy, in that one species of flagellin is sufficient to sustain motility. However, when that flagellin is FljJ cells are non motile. Distinct flagellar assembly checkpoints are utilised by bacteria in order to coordinate and couple flagellar gene expression to the assembly pathway. One of these checkpoints, hook-basal-body completion, is sensed and overcome by the secretion of a flagellar-associated secretion substrate. This mechanism results in a system switch to the export of proteins, which are needed for filament assembly. In C. crescentus the post-transcriptional regulators FlbT and FlaF have been implicated to function at this switch. As the flagellins themselves are being secreted we asked the logical question: are the flagellins involved in subunit feedback control of the regulation of filament assembly? In a bacterial two-hybrid assay, FljJ was the only flagellin able to interact with FlbT and FlaF. Furthermore, FljJ along with FlbT was found to interact with flagellin mRNA. In light of our data we a propose model for the regulation of flagellar filament assembly in C. crescentus and discuss the implications with respect to other flagellar systems

This thesis seeks to longitudinally assess the quality of public debate within the Twitter discussion of Brexit, undertaken through a mixed-method approach that combines quantitative and qualitative measures, as well as human coded and computational elements. To undertake this research, a corpus of N=450,000 tweets have been retrospectively gathered across four years, from February 2016 until February 2020, allowing my research to draw conclusions based upon trends across time. The theoretical framework is predominantly informed by Habermas's public sphere theory, which establishes the basis for my empirical research. Specifically, I argue that quality public debate should be free from polarisation and distortions in power relations, being an accessible space that promotes rational discussion that aids in the formation of informed public opinion. Within the literature review, I demonstrate that the EU referendum suffered from a deliberative deficit, which may have negatively impacted discussion on the platform and challenged the legitimacy of the 2016 referendum result. Through the empirical findings, I then highlight several issues that constrain Twitter's ability to facilitate healthy public debate within the context of the UK-EU discussion, with several measures pointing to a long-term decrease in the quality of the conversation. However, a number of the empirical findings are highly nuanced, which may reflect the complex nature of political discussion on the platform Twitter. In presenting the empirical findings, this thesis challenges previously held assumptions about the UK-EU discussion on Twitter, as well as contributes to new understandings regarding the quality of contentious political debate on the platform.

This study investigated the potential of undertaking sociocultural evaluation of data visualisation within the context of commercially driven data visualisation production. This study takes the stance that current approaches to evaluating data visualisation do not take into consideration the context within which they are consumed, as well as the external factors that can impact upon this consumption. As such this study drew upon the work of Seeing Data, in particular the work of Kennedy et al., (2016) in utilising the approach of Stuart Hall (1973) in understanding the encoding and subsequent decoding process of consuming data visualisations to develop a sociocultural approach to evaluating data visualisations. In doing so, this study utilised an interdisciplinary mixed methods approach, utilising both qualitative and quantitative data collection. In order to ascertain the existing forms of evaluation, and under which conditions they took place, a participant observation oriented ethnographic study of a commercial design studio took place over two visits. Drawing on the data collected and analysed from these visits, as well as the existing literature on the evaluation of data visualisation, a digital tool was developed through an iterative design process to capture sociocultural data from the consumers of data visualisation. This tool was developed using a participatory design approach, involving user testing and the input of data visualisation practitioners through a co-design session. The tool was subsequently evaluated using the approach of tool criticism, utilising a critical tool-walkthrough method to interrogate the suitability of the tool to capture sociocultural data as well as its place within the production process within the commercial setting. Findings from the commercial design studio revealed the complexities of undertaking sociocultural evaluation at scale within the constraints imposed by the resources available. These constraints resulted from the reliance on the existing conventions of data visualisation practice, and the availability of resources such as time, finances, as well as access to the end-consumers and data. Often these constraints arose out of the complex relationship with the client that had to be negotiated by the design studio and its practitioners. The tool received praise from the practitioners and proved successful in gathering data valuable to the production process. As such this revealed the potential for developing tools, and therein tool kits and novel approaches, that can be imbedded within the production process to navigate the constraints found within the commercial setting of data visualisation production.