Content Analysis

Content analysis is an analytic method used in either quantitative or qualitative research for the systematic reduction and interpretation of text or video data. Data can be generated from a variety of sources including (a) individual or focus group interviews; (b) responses to open-ended survey items; (c) text from social media; (d) printed materials such as research articles, newspapers, or books; (e) video-taped simulations; or (f) naturally occurring conversational events. It is also used in case study research. The aim of content analysis is to describe data as an abstract interpretation.

Use of content analysis as a research technique dates to the 1900s when it was used in communication research primarily to describe the quantity (frequency) rather than quality (meaning) of content contained in textual data. Since this early use, qualitative content analysis has gained popularity as a means to interpret data by identifying codes and common themes (manifest content) and then constructing underlying meanings (latent content). Content analysis is estimated to have been used as a qualitative analytic method in more than 3,000 research studies between 2005 and 2015 in such diverse fields as education, business, economics, social work, social science, and health sciences, including nursing, psychology, medicine, rehabilitation, gerontology, and public, environmental, and occupational health.

At least three distinct approaches to content analysis have emerged. These approaches differ in terms of study design, sampling decisions, and analytic strategies used, particularly how coding schemes are developed. The selection of approaches to content analysis largely depends on the research purpose and the availability of existing knowledge in the area of interest, particularly related models or theories. When existing knowledge around a phenomenon of interest is largely absent and the purpose of a study is to create knowledge, an inductive approach or conventional qualitative content analysis is appropriate where codes and themes are generated directly from the data.

When prior research or theory exists and the purpose of the research is to confirm, expand, or refine this existing understanding of a phenomenon, [Page 393]a more deductive approach or directed qualitative content analysis is appropriate using existing knowledge or theory to build the initial coding structure. When quantification of a specific content is desired, a summative content analysis approach is appropriate to identify and tally keywords or concepts.

As with any research method, sampling decisions are critical to meet study goals when using content analysis. Generally, sampling in a qualitative design seeks to maximize diversity of data around the phenomena of interest. Sample sizes may vary considerably when using content analysis depending on the research question. To understand a complex emotional event, researchers might conduct in-depth interviews with a small number of participants, while to understand what terms are used to describe a physical symptom, researchers might analyze written responses to an open-ended survey item from hundreds of participants. Using a directed content analysis approach, a researcher might purposively sample a particular group to refine or extend existing knowledge or theory about a particular phenomenon to a new population.

The development of the initial coding scheme and overall approach to coding differs depending on the specific content analysis approach chosen. With a directed content analysis approach, the researcher develops an initial coding scheme from existing theory or knowledge, using the data to modify or expand these codes. In a conventional content analysis approach, the initial coding scheme emerges from the data. With either approach, generally, it is helpful to first immerse oneself in the data to obtain a sense of the whole. Then data are coded through an iterative process. It is important to identify a consistent unit of coding, which might range from a single word to short paragraphs. Coding serves to reduce and condense the data based on its content and meaning. Finally, the relationships between codes are constructed by arranging them within categories and themes.

...