Qualitative data analysis using data displays

The amount of data generated in qualitative research can be difficult to manage. In this paper Tracey Williamson and Andrew Long discuss how the use of data displays can improve data management and also how the process can help to make the routes from raw data to research findings in qualitative research more transparent. Data displays can take several forms but share the benefit of helping to condense large amounts of data into more manageable forms. They can also help to convey information in a visually stimulating format where presentation time or column space may be limited

key words

* data displays

* qualitative data analysis

* methodological rigour

* shared governance

Introduction

One of the many challenges facing anyone undertaking qualitative research is how to organise and analyse the copious amount of data generated. These often take the form of pages of interview transcriptions, memos and field notes, sometimes together with observational data, that are cumbersome and time-consuming to manage (Morse and Field 1996). Many approaches to getting to grips with the volume of data are evident in the literature. These range from the use of computer packages such as N-Vivo[TM], thematic analysis (for example, using the 'framework' approach of Ritchie and Spencer 1994) to the classic paper-and coloured pens approach, with data placed in separate piles in a large work space. Between these extremes lies the structured approach outlined by Miles and Huberman (1994) which involves a three-pronged strategy: data reduction (the process of selecting, abstracting and transforming the basic data, that is, its initial coding and search for themes): data displays: and conclusion drawing and verification.

This article explores the second part of their approach. The aim is to provide an insight into the range of possible data displays and critically to examine their use as aids to qualitative data analysis. The paper draws on our experience of using this approach during a doctoral study of group decision making in shared governance (Williamson 2003). Shared governance, a relatively new concept in the UK, concerns the empowerment of healthcare staff to contribute to decision making that will in turn influence practice and the corporate agenda. The study sought not only to identify factors affecting decision making in shared governance, but also to understand the processes between these factors that led to effective decision making.

Data were generated from over 200 hours of participant-observations of shared governance council meetings and a range of individual (31) and focus group (three) interviews, plus secondary data such as strategy documents and minutes of meetings.

What are data displays?

Miles and Huberman (1994) describe a data display as 'an organised, compressed assembly of information that permits conclusion drawing and action'. Data displays can take a number of forms. Extended text itself is a data display (Box 1). Others include matrices, graphs, charts and networks. These can be in the form of basic data displays, such as time-ordered matrices, of condensed versions of data spread out over many pages. Or it may be advantageous to stack these to...