Break up the raw data derived from interview transcripts, observation notes, or other sources into manageable parts, elements, or units, to identify types, classes, sequences, and patterns. Such break-up reconstructs data and lends clarity to the study. Pair the available raw data by coding, identify themes that describe the phenomenon under study, and assign the coded data to the appropriate theme.
Codes are surrogates for textual data, summarizing and synthesizing observations contained in a set of raw data. Researchers use codes to emulate distributional analysis and hypothesis testing, and to group or categorize otherwise discrete events, statements, and observations into distinct classes. The method of coding extends to assigning numbers, symbols or objects, or any other form of representation.
Thematic analysis or identification of themes to categorize the coded data resembles sorting a box of buttons, which people group according to size, number of holes, color, or type. Identify themes and categorize data to such themes similarly. When creating categories, include actual label or theme, define the theme, describe how to recognize a theme, describe any qualifications or exclusions to identifying themes, and provide examples that lend clarity and eliminate possible confusion.
Classification of coded data or assigning the coded data to an appropriate theme is similar to working on a jigsaw puzzle. Just as one starts by sorting the different pieces of a jigsaw puzzle based on the objects represented in the puzzle, classify the coded data based on their similarities. Two standard classification systems are typology or straightforward classification into groups, and taxonomy, ormultiple level taxology with super-ordinate and subordinate categories. Some sets of data classify esily, while other sets of data remain difficult to classify.
Many software programs assist data coding, thematic analysis and classification. Coding and retrieval programs help the researcher retrieve text segments from the data set and code them, and theory-generating programs identify relationships between coded categories theoretical explanations.
The process of coding, identifying themes, and classification is subjective, and success depends on the skill and experience of the researcher.