Discourse analysis: theoretical and historical overview and review of papers in the Journal of Advanced Nursing 1996-2004.

AIM The aim of the paper is (1) to offer an overview of different theoretical approaches to discourse analysis and (2) to review papers published in the Journal of Advanced Nursing from 1996-2004 in which discourse analysis is identified as a method of data analysis. BACKGROUND Discourse analysis offers rigorous approaches to analysing naturally occurring talk and texts. Forms of discourse analysis have developed across broad theoretical bases. Such development has created challenges for researchers wishing to adopt this methodology and readers wishing to evaluate the quality of discourse analytic work. METHODS First, key documents which describe the theoretical range of discourse analysis are used to provide (i) a comprehensive overview of the approach, (ii) the identification of categories of discourse analysis and (iii) minimum criteria for determining if an individual paper can realistically claim to be adopting discourse analysis. Secondly, an electronic search followed by hand search of the Journal of Advanced Nursing full-contents between 1996 and 2004 was undertaken. The papers were grouped into the types of approach identified in i, and evaluated to see whether they met the 'minimum criteria' also identified in i. FINDINGS The search of Journal of Advanced Nursing revealed 24 papers where the authors stated that discourse analysis was among the methods or was the sole method of data analysis. The majority of the papers cluster around critical approaches to discourse analysis. Only a few approach discourse analysis primarily as analysis of conversation. Some papers are excellent, while others offer analysis that bears little resemblance to any form of discourse analysis. CONCLUSIONS A strategy for improvement could include more rigorous attention on the part of those practising discourse analysis to methodology and the key features that differentiate the different approaches to discourse analysis from other qualitative methods. Authors should include sufficient detail of their approach.