This website is archived and is not maintained.
For up to date information on eResearch, go to our NEW Manchester e-Research Centre Site, MeRC http://www.merc.ac.uk

Using Text Mining for Frame Analysis of Media Content

Using Text Mining for Frame Analysis of Media Content

 

Project Start:	1 February 2008
Project End:	31 July 2009
Project Partners:	National Centre for Text Mining (NaCTeM) & ESRC Centre for Research on Socio-Cultural Change (CRESC)
Funding Programme:	JISC support of research committee (JISC), e-Infrastructure Programme, e-Research Theme

Summary of the Project

Text-mining technologies offer the opportunity of processing large amounts of textual data systematically, reducing human errors, and saving time. They have the potential to at least partly automate the generation of frames (the process at the heart of frame analysis) to a greater extent than possible using current Computer-Assisted Qualitative Data Analysis Software (CAQDAS) packages. The project, as a use case of the JISC-funded ASSIST project , will not only produce advanced ICT tools for enabling social scientific research, but also help establish the foundations for the wider adoption and sustainability of NaCTeM text-mining services.

 

Background

Frame analysis has been widely adopted for investigating how texts are framed in a certain way to shape the perceptions or opinions of the information's recipients. Although computer-assisted qualitative data analysis (CAQDAS) packages are available to manage and manipulate textual and/or multimedia data, they are not sufficiently advanced to automate the interpretive work of coding that lies at the heart of frame analysis, nor do they support complex retrievals that need to cope with language variability such as synonymy and polysemy. This project will explore the usefulness of text-mining techniques for the analysis of large media corpora. It builds on the Automatic Summarisation for Systematic Reviews using Text Mining ( ASSERT) project.

 

Aims and objectives

This project, in collaboration with the ESRC Centre for Research on Socio-Cultural Change (CRESC) and National Centre for Text Mining (NaCTeM), aims to illustrate how text-mining technologies might advance frame analysis in social science research. The project has two objectives: 1) customising ASSERT's tools for application to frame analysis of newspaper text; 2) providing a use case to extend awareness and promote adoption of text mining across all social science disciplines.

 

Project methodology

To achieve the above objectives, the project will

-  investigate frame analysis practices to define initial user requirements for text mining tools;

-  use an iterative process based on design, rapid prototyping, evaluation and refinement, to customise the ASSERT suite of text mining tools for the qualitative research community;

-  establish an evaluation framework which may be used in new applications of text mining tools;

-  investigate potential barriers to adoption, sustainability issues and establish responses such as user training and support;

-  document the application of text mining tools for media research as an e-framework use case.

A domain expert-led frame analysis of the newspaper coverage on the introduction of ID cards in the UK has been essential for understanding the mainstream frame analysis practices and for identifying user requirements. The domain experts involved in this project have also participated in the tool development process, interacting with the developers regularly.

Anticipated outputs and outcomes

The deliverables of this project are:

Software Outputs

-  A frame analysis demonstrator based on a customised version of current ASSERT tools;

Non-Software Outputs

-  Documented processes and workflows;

-  A case study demonstrating the use of ASSERT text mining tools in social sciences;

-  Evaluation framework;

-  Use case

-  Barriers report

-  A final evaluation report summarising major findings, lessons learnt and the impact of this project's approach to facilitating effective frame analysis.

 

Technology / Standards used

In the future, the aim is to integrate these text mining tools with other e-Research tools for linking, processing, managing and sharing multiple forms of social scientific data. There is thus a need for greater coherence in development, and for a map of what has been developed and the standards and specifications that underpin them. This information will enable a strategic approach to planning programmes of development, and would provide institutions with information on what is available and ready for adoption and mainstream use. As such, we will commit to open standards, and encourage interoperability and tool integration.

 

Related Publications

Pieri, E. (2009). "ID cards: a frame analysis of the debate in the UK printed media." Paper presented at the Politics and Web 2.0 Research Methods Workshop, Manchester 19-20 February 2009.

Panel on "Innovations in Methods in Media and Communication Studies" at the Media, Communication and Cultural Studies Association (MeCCSA) Conference, Bradford, 14-16 January 2009. (pp. 37-39 in the conference abstract book)

Lin, Y.-W. (2009). "Some Methodological Thoughts on Using Text Mining Techniques for Frame Analysis of Media Content." Paper presented at the panel on "Innovations in Methods in Media and Communication Studies" at the Media, Communication and Cultural Studies Association (MeCCSA) Conference, Bradford, 14-16 January 2009.

Pieri, E. (2009). "The introduction of ID cards in the UK: A snapshot of the debate in the press." Paper presented at the panel on "Innovations in Methods in Media and Communication Studies" at the Media, Communication and Cultural Studies Association (MeCCSA) Conference, Bradford, 14-16 Jan 2009.

Pieri, E. (2008). "The National Identity Scheme and the introduction of biometric identity cards: The debate in the press." Paper presented at the "Text Mining and the Social Sciences" Workshop at the 4th International Conference on e-Social Science, Manchester, 18-20 June 2008.

Workshop: Text Mining and the Social Sciences, the 4th e-Social Science Conference, Manchester, 18-20 June 2008. Slides are available here.

Core Project Members

Prof. Peter Halfpenny, National Centre for e-Social Science, University of Manchester (Principal Investigator)

Dr. Farida Vis, ESRC Centre for Research on Socio-Cultural Change, The Open University (Co-investigator)

Dr. Yuwei Lin, National Centre for e-Social Science, University of Manchester (Project Manager)

Elisa Pieri, National Centre for e-Social Science, University of Manchester (Principal Researcher)

Dr. Davy Weissenbacher, National Centre for Text Mining, University of Manchester (Principal Researcher)

Prof. Peter Golding, Pro-Vice-Chancellor (Research), Loughborough University (Advisor)

Dr. Thomas Koenig, QUIC node of the Nationa Centre for Research Methods, University of Surrey (Advisor)

The ASSIST Project 

Dr. Sophia Ananiadou, National Centre for Text Mining, University of Manchester (Principal investigator) 

Dr. Brian Rea, National Centre for Text Mining, University of Manchester (Project Manager)

Dr. Davy Weissenbacher, National Centre for Text Mining, University of Manchester (Principal Researcher)