The Coding Manual for Qualitative Researchers, Chapter 1: An Introduction to Codes and Coding

Text Exploration

In this chapter, the author introduces the idea of codes, and related concepts (like categories, lenses, themes, and filters), providing practical techniques and examples of the coding process and output.

Codes and coding are contentious, and the author describes some of the conflicting perspectives on the role of this activity in social science. Strauss indicates that “the excellence of the research rests in large part on the excellence of the coding” while Packer describes that coding is “impossible in practice.” This book emphasizes the value of coding.

A code is a word or phrase that assigns an attribute to a piece of data. Coding is cyclical, and author recommends beginning to code in parallel to the research process—during data collection, not after the fieldwork is done. In all cycles of coding, coding can be as large as a paragraph of text. Codes translate data; the translation process is not an exact science, and is “primarily an interpretative act.” Codes help identity patterns, and qualitative researchers look for these patterns as they learn how people make the world more understandable; they identify routines, rituals, rules, roles, and relationships. Patterns indicate similarity, difference, frequency, sequence, correspondence, and causation.

The author offers a diagrammatic representation of how data leads to codes, codes to categories, categories to themes and concepts, and concepts to a theory or assertion.

Codes and patterns codify ideas; “to codify is to arrange things in a systematic order” and coding is a process of codifying. Synthesis, then, codifies and transitions from coding to categorizing. Qualitative inquiry focuses on language and images, and coding forces a researcher to explore those ideas.

There is a difference between codes and themes. Themes are an output of coding, but are not themselves coded. A category is a phrase describing something in data that is explicit, but a theme is something that is more subtle or tacit.

There is conflict amongst scholars about how much to code. Some feel that everything should be coded, while others capture only a small portion of material. Detailed coding can be thought of as “splitting,” in that individual parts of a sentence may have different codes, while “lumping” describes coding only paragraphs or chunks of text. These lumps are sometimes referred to as stanzas. Some recommend developing 50-300 different codes; others aim for a smaller number, ultimately organized into 15-20 categories.

The codes are then stored in a codebook, which is a compilation of the codes, their descriptions, and an example; more specifically, the codebook contains a short description, a detailed description, inclusion criteria, exclusion criteria, examples, atypical examples, and “close, but no” examples. The author recommends that a dissertation include the codebook itself as an appendix to the study, while published papers rarely need to offer that detail.

On a group research project, team members may code together with the goal of achieving “intercoder agreement” or “interpretative convergence.” In these cases, the goal is to align 80-90% of coded material across coders. The author, however, prefers to code alone.

There are a variety of researcher traits, or qualities, that the author has identified as being important for researchers who take on this technique. These include being organized, having perseverance to complete the tedious and frustrating process, dealing with ambiguity and being flexible, being creative, as social science has “a lot of art” to it, remaining honest, and having and leveraging an extensive vocabulary.

Some feel that coding is reductionist, mechanistic, instrumentalist, and distances a researcher from the data; the author describes that “nothing could be further from the truth” and that coding should force reflection.