ESRC Methods Festival, 1 - 3 July 2004

‘Software for media analysis / Social networks’ symposium (2^nd of July. 14.00 ─ 15.30)

Software in media analysis:

The Uses and Misuses of Qualitative Data Analysis Software

in Qualitative Research.

Michael Billig, David Deacon, Peter Golding, Thomas König, Katie MacMillan, Department of Social Science, Loughborough University

This paper will address the evolution of analysis of media content and will contrast its development with the scale and nature of theoretical and methodological work in relation to cultural consumption in particular. The broad aims of the project will be outlined to set in context our work on the application of software to media analysis and of comparative research into media output.

We will examine the kinds of materials researchers work with in analysing the media, and will appraise the value and reliability of commercial archives.

Analyses using qualitative data analysis software have become more common in social scientific research in recent times, but have, as yet, been little used in media research. We will evaluate qualitative data analysis software and its contribution to various methodologies ─ including traditional content analysis, frame analysis, and discourse analysis. Our discussion will be illustrated with examples from our analysis of media content on topics including the UN Weapons' Inspections in Iraq in the build up to the second Gulf war, and controversies about Anti-Semitism talk in political speeches across a number of countries.

Content Analysis and Media Archives

• The first step in using qualitative data analysis software for media analysis is to save texts in digital form to allow them to be imported into the software. A range of media data bases are now commercially available that provide access to media content in this format. Moreover, their search and retrieval facilities apparently provide the opportunity for instantaneous, reliable and extensive surveillance of media archives, whether to identify raw data for further analysis or to draw immediate inferences about general trends in media coverage.

• Although studies utilising these kinds of ‘push button’ content analysis have increased over the recent years, there has been little methodological reflection about the strengths and weaknesses of using these information resources in these ways. This part of the presentation will address these questions and present the findings of our assessment of the LEXIS NEXIS News Archive, an on-line newspaper archive that has achieved a position of considerable market dominance.

• The specific issues raised include asking which kinds of content analysis can be conducted through this form of electronic searching (and which cannot); what recommendations there are for good research practice when using these digital data bases; and what the methodological implications are when accessing press content solely in digital text format.

Qualitative Data Analysis Software and Frame Analysis

• This section will look at qualitative data analysis software, and its suitability for textual analysis based on searching and interpretative coding strategies ─ such as so-called Qualitative Content Analysis, and some forms of Frame Analysis.

· We will discuss how interpretative codings ─ which allow the coding of words and phrases that are meaningless when stripped out of context, but central to a contextualised understanding of discourses ─ can efficiently be performed using some QDAS packages. We will examine how QDAS can be used in frame analysis to go beyond the (relatively) more common use of unique keywords to identify frames in larger corpora of data.

· We will explore the extent to which such processes may enhance the validity and reliability of research results.

Qualitative Data Analysis Software and Discourse Analysis

• There are a variety of different forms of discourse analysis – some more textually and linguistically based, while others are more pragmatically and interactionally based. However, all forms of discourse analysis should involve analysing how the discourse in question is constituted and what it is being done rhetorically. Unfortunately, what often passes for discourse analysis in the social sciences often involves a minimal analysis of the discourse as such. Some basic failures to engage in analysis will be discussed, such as tendencies to summarise or to categorise rather than analyse.

• The above points will be related to the issue of using computer packages to conduct discourse analysis. There are a wide variety of software packages for qualitative analysis designed to do the practical tasks of sorting, storing, editing and coding, traditionally done by hand. There are, however, very few discourse studies firmly rooted within a particular discipline that use (or discuss) QDAS. The exceptions to this are Corpus Linguistics, and, to a certain extent, Conversation Analysis ─ the former lends itself readily to the search, count, and code facilities of computer programmes; the latter is more concerned with the technical value of the software functions, and the extent to which it can allow audio and video recordings to be played simultaneously with transcriptions of conversation.

· Qualitative data analysis programmes do not, however, on their own produce the level of analysis that is required for discourse analysis. One reason is that all forms of discourse analysis require that the material to be analysed should be understood in relation to its particular discursive, interactional or rhetorical context. This means that its particularities must be studied and it is not sufficient merely to consider these particularities as instances of something more general. For this reason, discourse analysis cannot simply employ a universal set of procedures, which could be formalised into a computer package. Instead, discourse analysis always poses new problems that in their turn make new demands upon the analyst.