Authors of this page: Graham R. Gibbs¹, Celia
Taylor¹ and Ann Lewins²
Affiliation: ¹University of Huddersfield and
Date written: 14th Nov 2005
Coding involves identifying words/phrases/lines/sentences/passages of text in a document or an image or part of an image that represents an idea or concept. This is then linked to a named code that represents that idea or concept. This shows that it shares the characteristics indicated by the code and/or its definition with other similarly coded passages or texts. All the passages and images associated with a code can be examined together and patterns identified. Programs differ in how they support coding (including in vivo coding, and code creation) and how they manage to show coded segments in context.
Codes can be arranged non-hierarchically (free codes) as a simple list. This enables nodes to be created without having to worry immediately about how they relate to other codes. The codes can also be arranged as a hierarchy (tree) with a branching arrangement of codes. Ideally, child codes in a tree relate to their parent codes by being 'examples of...', or 'contexts for...' or 'causes of...' or 'settings for...' and so on. Most CAQDAS packages allow flat and tree (hierarchical) coding, but some software, though they principally allow for flat coding, do allow you to make collections of codes or draw connections between them that allow for types of hierarchical approach.
The function in some CAQDAS programs to code the results of a search for words or phrases in the text itself, e.g. in MAXqda and Atlas.ti this is an option. In NVivo and Nud.ist by default, searches produce a new node.
Variables are a general characteristic or quality, often shared by all cases in the study, such as their social class, gender, or role etc. Variables have two or more values and usually each case or data file is assigned just one value for each attribute/variable. Values may be numeric or categories.
A memo is a document used to reflect on the analysis for instance the meaning of data or codes (Weitzman & Miles, 1995). Memos may be separate documents, attached to the text or codes (especially in a CAQDAS program) or collected to form a research diary.
Search is one of the core functions of CAQDAS and includes both lexical searching (searching for words and phrases repeatedly in text) and code searching. In the case of code searching, what is found by the search is passages of text that are coded (or not) in specified ways and that are related to other coded passages in specified ways (e.g. they overlap, they are coded by both codes).
Text (lexical) search
A search tool that allows you to search for words, terms and phrases. The search can be done for a single word or a line of words (string) that are related in some way. A core activity of the software is text searching or lexical searching. Programs support this in a variety of ways. There are some fruitful approaches to using word searches. Most CAQDAS software (except the Ethnograph) provide word or phrase searching tools. If, however, what is required is a total word indexing or listing or counting across all or some of the data. In some CAQDAS programs it requires add on modules (such as MAXDictio for MAXqda, or Wordstat for QDAminer). In ATLAS.ti there is a built-in, basic word frequency tool. In other cases separate programs such as WordSmith, Concordance and Conc can be used alongside CAQDAS programs.
Many programs now have functions that allow searching for combinations of coded text. Usually these include both Boolean (and, or, not) and non-boolean or proximity (near, after, overlap) combinations. These are very powerful facilities. An important use for the results of code searching is in various forms of comparative analysis.
A Boolean search uses a combination of AND, OR, NOT (Weitzman and Miles, 1995). For instance one may search for text that matches the words 'luck' And 'chance' or for text that is Not coded at the code 'Individual interview'. Databases (for example, library catalogues) and CAQDAS programs commonly support such searches.
Code 1 AND code 2
Code 1 OR code 2
Code 1 NOT code 2
When to use
To find out whether people who talked about being depressed also talked about living alone or being lonely.
The non-Boolean searches are co-occurrence, sequence and inclusion of coded passages of text. A search to find whether a combination of words or themes occur NEAR, FOLLOWED BY, OVERLAPS, within a set area of text. Near, means text near to each other. Follow, means text that follows each other. Overlaps, means that coded text overlaps each other (Weitzman and Miles, 1995). The way such searches are described and defined in each software may be slightly different.
Code 1 NEAR Code 2
Code 1 FOLLOWED BY Code 2
Code 1 OVERLAPS Code 2
When to use
To find out whether people who talked about being depressed also mentioned social support at the same time.
Models and networks are graphical models to show conceptual, theory building relationships between codes, memos and documents. They are diagrams designed to represent the relationship between key elements in a field of study. Models may be predictive, causal or descriptive, and may be discursive, mathematical or graphical. The models help clarify and test ideas (Weitzman and Miles, 1995).
Although anathema to some researchers, many researchers want to integrate related quantitative data with their qualitative analysis. Researchers increasingly seek to include a mixed methods dimension and quantitative data can be imported into CAQDAS packages as a way to organise qualitative data, to assist in triangulation, comparative enquiry and searching. Less often utilised is the ability to export qualitative analyses as quantitative data. Quantitative methods can also be used to support qualitative enquiry as exploratory approaches.
Hypertext is a link to a different part of the database for instance between codes, memos or documents (Weitzman and Miles, 1995). The links can also be between whole files or between particular places in the text. The linkages are used to show logical and meaningful connections between the passages and their content. The world wide web is an example of a hypertext. Several CAQDAS programs support such linkages to different degrees.
In a CAQDAS program this is usually a rich text file which the user may create from within the program. Its contents may record various kinds of data generated by the program, for example the results of a search. The generated output file can then be read in other applications such as Wordpad or Word.
Weitzman, E. A. and Miles M. B. (1995). Computer Programs for Qualitative Data Analysis: A Software Source Book. Thousand Oaks, Sage.