Constructing the first version of the initial template
Affiliation: University of Huddersfield
Date written: 12th January 2009
- In some approaches to thematic analysis, such as template analysis, the researcher can construct a coding scheme or coding template before starting to examine the transcripts of interviews or ethnographies.
- There are various sources of ideas for initial coding templates: the interview schedule, the researcher's own experience, the researcher's professional knowledge and the transcripts themselves.
- Common problems with coding schemes or templates that most researchers experience are that the codes are simply descriptive - they go no further than simply summarising the respondents own words - and there are too many codes in the scheme or template to be useful in an analysis.
Coding is at the heart of much qualitative analysis and in order to code the texts they are analysing, researchers have to construct a coding scheme or template. This is a list of the codes that can be or have been applied to the texts. The lists may be hierarchical, that is to say there are high level codes - sometimes referred to as themes or categories - under which are grouped several sub-codes that represent examples, types, settings, causes etc. of the high level codes.
In some approaches, most notably grounded theory, the advice is to develop these coding schemes by reading the texts so that the themes are, by definition, grounded in the textual data. However, commonly and as recommended by other approaches such as template analysis, framework analysis and interpretative phenomenological analysis, as well as by those adopting a less rigorous form of grounded theory, researchers construct an initial, a priori set of codes or coding scheme which is then progressively amended and modified during the analysis and coding of the transcripts. There can be a variety of sources of inspiration for these initial coding schemes or templates: topics on the interview schedule that was used, themes from the research proposal, ideas from the literature on the research topic, issues gathered from professional experience or personal experience of the field and the researcher's own hunches about the issues. These are quickly augmented by ideas and themes derived from the direct examination of the texts being analysed.
Two common problems that researchers encounter when they start coding, are that they create too many codes and thus end up with an unwieldy and unhelpful coding scheme or template, and that their codes are too descriptive. These problems are often interlinked. Descriptive codes tend just to summarise the various ways in which respondents have talked about the issues and tend to use their terms or words. At the start, researchers want to capture the diversity of ways that respondents experience the world and talk about it and they do this by creating codes for each example. The challenge it to reduce the number of codes by replacing them with more analytical ideas that encapsulate the underlying commonalities of several codes. This often means using terms that the respondents did not use, but which nevertheless capture the essence of what they were experiencing.