Web Sphere Analysis: A (Very) Brief Overview

Modern public spheres are based on mediated forms of communication that provide a shared catalogue of references for social collectives. The world that we know, including our broader cultural, economic, social, and political environment, becomes to large parts accessible through distorted representations distributed via mass media and the Internet. Especially web communication continues to gain in relevance as a crucial, highly personalised and customisable source for information about social reality.

Web discourses hence remain fresh and relevant subjects for research on public communication in networked, highly mediatised societies. But not only researchers in media and communications have an interest in theorising and empirically analysing digital public spheres; policy-makers, political consultants, and media monitoring agencies have recognised the relevance of the Internet as a space of resonance to political, economic, cultural, and social developments.

However, the precise identification and evaluation of web debates is a considerable methodological challenge. A very fruitful approach to define and approach these extremely dynamic communicative contexts provides the web sphere perspective as proposed by Schneider and Foot (2006). Their methodological proposal to understand web discourses as a condensation of related online communication, i.e. content, enables the critical observer to identify, analyse, and assess digital public spheres efficiently and link them to the underlying social and political currents that sparked their formation. This post provides an (extremely) short summary of its core ideas.

What are Web Spheres?

A web sphere is a collection of related online content that focus the same set of issues or events; the respective content thus shares a common context and is (potentially) held together by referencing and/or hyperlinking. A web sphere can integrate different kinds of online platforms and formats, which highlights the degree of interconnectedness that web content can display. For example, the war of frames/”digital words” between Islamic extremists and their opponents is not limited to one particular social media platform (e.g. Facebook) but includes the whole range of accessible web technologies.

Triggers for the formation of a web sphere are often irritations in everyday politics or society, such as scandals, disasters, terrorist attacks, but also seemingly trivial issues like celebrity news or the colour of a dress. One can basically differentiate between two general types in this respect: firstly, there are web spheres that are somewhat “predictable”, i.e. one can expect they will probably emerge in the context of a planned, ritualised event. Examples are web discourses on the next FIFA World Cup, the next general elections, or the next Academy Awards. Secondly, there are web sphere that emerge erratically as immediate “real-time” reactions to unforeseeable, ad-hoc developments and events. Accidents, disasters, unplanned revelations are often catalysts for an unscheduled torrent of online communication that potentially condenses into a web sphere. The sudden and still unsolved disappearance of an Malaysian Airlines plane in March 2014 is one such tragic event.

Web spheres can also differ in their degree of durability: some may vaporise as quickly as they formed, while others may persists over longer time periods. The issues that determined their emergence, i.e, their content, as well as the set of participating communicators, i.e. their underlying networks, are the most important factors that influence their duration.

How to Analyse and Evaluate Web Spheres?

Due to the virtually unlimited amount of web sources it is very difficult to define the actual borders of a web sphere. In fact, any meaningful empirical investigation is inevitably limited to a mere snippet of a potentially much larger web discourse. It comes all down to the general problem of sampling and representability of findings in online media research. However, these limitations have always affected analyses in media and communications to one degree or the other. When deciding what content is considered as part of a web sphere, it is absolutely crucial to explain its assumed relevance and to outline the limits of sampling.

Once a set of sources has been defined for analyses, it is recommendable to archive/store the respective websites (urls) with precise information on their origin, date of publication, authorship etc. for documentation. Online content is extremely dynamic and it can become very difficult to retrieve the original content after longer time periods.

The next step is to decide what the web sphere analysis is exactly focusing on; potential research questions can aim for demographic/ethnographic factors, networking patterns, and discursive practices (e.g. framing). For example, in my analysis of the EU crisis web sphere I combined frame- and network analyses to reveal how the Eurozone crisis was perceived from different cultural-political perspectives and what the social composition of communicators looked like. Data for both analyses was collected via a multi-step content analysis.

The in-depth screening of a web sphere in terms of its content and sociological properties therefore depends on the actual research interest and can be achieved through qualitative and quantitative content analytical methods.

In short, a web sphere may include the following steps:

  • Definition of the social phenomenon that causes online debates (e.g. a political development or cultural event)
  • Identification, sampling, and archiving of relevant online content (e.g. snow ball sampling, combination of non-probalistic and random sampling)
  • Qualitative and quantitative analysis of the web sphere’s content and/or sociological factors (e.g. frame analysis, discourse analysis, network analysis etc.)
  • Presentation of results (and potential predictions for future developments)

The public sphere – seen as a highly differentiated and dynamic network of media-based discourses  – experiences constant transformations; Internet technology is a driving force in these processes and understanding the structure and logic of web discourses is absolutely crucial for researchers and practitioners in public (political) communication. In this respect, the web sphere model provides a flexible, easily customisable as well as expandable methodological approach for comprehensive analyses – which is a starting point for grasping the complexity of public discourses in the networked society.

References:

Schneider, S. M. and Foot, K. A. (2006): ‘Web Sphere Analysis. An Approach to Studying Online Action’, in Hine, C. (ed.) Virtual Methods. Issues in Social Research on the Internet, Oxford and New York: Berg.

Image courtesy of http://hubblesite.org/gallery/album/

Convergence in the News Media Discourse (Online) on the EU Crisis shown in Network Graphs

Here are a few graphs that I’ve created with Gephi, a very powerful tool for network graphs; the software is open source and you can download it for free. It might seem a bit intimidating to people who have absolutely no experience with network analysis software but there are tons of helpful tutorials out there, especially on YouTube. I will compile a list with my favourite ones in a follow-up post soon. It’s also good to have some knowledge about either Excel or SPSS (best both), as you will need data tables for calculating the network graphs (in CSV).

I use Gephi for the visualisation of  hyperlink patterns in online content but also for mapping relations between political actors, organisation, and, as in the present case, nation-states. For example, the  graphs below show what countries were mentioned and with which other nation-states they were most frequently contextualised/linked to in EU crisis content from three European news media websites (Guardian, EKathimerini, and Spiegel Online, click to enlarge). I extracted the data from over 13.000 online articles published between March 2011 and March 2013.

A node’s size indicates how often a specific country was named; the bigger it is, the more frequent it was mentioned or referred to in the sample. The lines (also called edges) show what countries were named together in the analysed content and their “thickness” indicates how often that occurred. For example, Germany and Greece have a very thick connection across all platforms, which tells us that they were particularly often linked to each other in the EU crisis discourse; considering Germany’s dominant role in the EU bailout negotiations for Greece, this is little surprising.

Guardian Online Nations Network EU Crisis 2011-2013
Guardian Online Nations Network EU Crisis 2011-2013
EKathimerini EU crisis graph
Ekathimerini Nations Network EU Crisis 2011-2013
SPON Nations EU Crisis
SPON Nations Network EU Crisis 2011-2013

Still, the similarities across the four different European news media sites indicate significant tendencies towards discursive convergence in the EU crisis debate. The same can be said of Le Monde for which I have created the same kind of graph. Each platform seems to put particular emphasis on its own national context (UK, Greece, Germany, France etc.) but in sum the networks between countries  that are somehow involved in the EU crisis discourse look very similar. This type of graph thus allows to analyse the set of entities that are involved in a particular political discourse and enables a more detailed evaluation of how their relations to each other are portrayed in media content.

SPSS PASW Tutorials

I finally managed to finish my MA dissertation on political blogs in the UK. As many of my observations and hypotheses are based on quantitative data, I used SPSS/PASW to calculate comparable statistics. It is one of the most powerful and versatile pieces of software for such purposes. However, it is not very easy to use and might repel some users at first.

Various books on the issue exist, which can significantly differ in their quality: Some are endeavoring to explain the very complex statistical mathematics step by step, to make them easily replicable. Others presume that the reader already has an elaborated knowledge of statistics and the formulas as well as terms in use. During my research I read through both kinds of books and I will list here two very useful introductions:

Field, Andy P. (2009) Discovering Statistics Using SPSS. Los Angeles & London: Sage.

This very comprehensive monograph explains everything about statistics in SPSS from the very basics. Field uses a comprehensive language, lots of illustrative examples, and succeeds in explaining the prosaic issue in an entertaining, humorous way. He elucidates quite eloquently the most important mathematical operations behind each step before he shows how to implement a certain type of statistical processing via SPSS/PASW. The ideal handbook for every person who wants/needs to deal with this program.

Bryman, Allan/Cramer, Duncan (2005) Quantitative Data Analysis With SPSS: A Guide for Social Scientists. London: Routledge.

As the title suggests, this book introduces SPSS for sociologists. Hence, it is applicable to certain branches of communication- and media studies, too. Though not fully ‘up-to-date’, the explanations for using SPSS in specific social science research projects are clear and convertible. Cramer and Duncan’s tone is more sober and less casual than Field’s style of writing. Nevertheless, this book is far easier to comprehend for beginners and non-staticians than many of its counterparts.

Moreover, there are various online sources. One notable example are the video tutorials provided by Central Michigan University (CMU) – these clips show beginners step-by-step how to use the software and solve statistical problems. Check them out here.