The content analysis is a widely practiced technique in qualitative inquiry. Although it originates in communications research, it also has a long history of use in journalism, sociology, psychology and business (Neuendorf 2002). As such, it is a generic name for a variety of means of textual analysis that involves comparing, contrasting and categorising a corpus of data, including now both numeric and interpretive means.
According to Gray (2014), content analysis involves the making of inferences about data (usually text) by systematically and objectively identifying special characteristics, such as classes or categories, within the data. In an attempt to achieve a measure of objectivity, the process allows for the creation of specific criteria for selection in advance of the data analysis process. These categories are often derived from the theoretical framework and brought to the empirical data and not derived from them (Flick 2009). In this study, the categories were derived as informed by the conceptual framework developed in Chapter 3.
Lasswell (1971) has described content analysis as a technique that emphasizes the quantification of “what” a message communicates and presents a classic formulation: WHO says WHAT to WHOM with what effect? “Who” refers to the source of information, the “why” refers to the codification process, and the “how”, refers to the communication channel, and the consequences or effects that the “receptor of the message” has (Holsti 1969). This was important in this study because the aim of the study was to determine the effects that the respective stakeholders have on the SMEs or the consequences of non-compliance.
4.7.2.1 Types of quantitative content analysis
There are two generic content analysis categories: (1) conceptual analysis and (2) relational analysis. Conceptual analysis is concerned with the examination of the reoccurrence of particular concepts in a text, while relational analysis builds
145
on conceptual analysis by examining the relationships among concepts in a text (Wilson 2011). The primary reason for the choice of content analysis over other qualitative method was its ability to quantify qualitative data. Therefore, in reference to Section 4.2, page 119 a conceptual analysis was used for this study. For this reason, it is explored in more depth below.
4.7.2.2 Conceptual content analysis
This is where a concept is chosen for examination and the number of times it occurs within the text being recorded. That is the content is coded for certain words, concepts or themes and the researcher make references based on the emerging patterns (Wilson 2011). In this study, the words coded included different stakeholders, stakeholder roles, current sustainability practices, sustainability transfer strategies, sustainability drivers and barriers, mechanisms for averting the barriers among others derived from the research framework.
4.7.2.3 Conceptual analysis methods
The method begins with the identification of the research question and the choice of the samples. In this case, it was the identification of research sub-questions. This is followed by the coding of texts in content categories. The codification process is essentially a selective reduction. Thinning out the text in categories, consisting of words, a set of words or phrases the researcher may put an emphasis, and to code, specific words or designs that are indicators of research questions (Rossi et al. 2014).
4.7.2.4 Steps for conducting conceptual analysis
These steps involve coding texts or a set of texts, which in this study were transcript interviews.
146
Table 4.3: Steps for conducting conceptual analysis
Steps Description
Establish the level of
the analysis The researcher decides the level of analysis. This may be coding a simple word or a set of words or phrases. Decide how many
concepts to code The researcher decides on the number of different concepts that will be coded. This involves the development of a pre-defined concept and categories. Thereafter, the researcher has to decide on how much flexibility will be allowed in the codification process. The determination of how many concepts or set of concepts allows the researcher to focus on specific points. Flexibility allows the inclusion of new materials for codification that may be important for the final result.
Deciding if concept codification will be made by existence or frequency
The researcher has to decide whether to codify only the existence or the frequency. The number of times that a concept appears in the text may be an indication of the importance of the concept.
Deciding how concepts will be distinguished
The researcher must decide on the level of generalization of concepts, i.e. whether the concepts will be coded exactly as they appear or may be counted as equals even if they appear in different forms. That is, whether the words mean the same thing or if the meanings are radically different. Develop rules to code
texts The development of a set of rules helps the researcher to ensure that the encoding of concepts is consistent throughout the text, and always the same way.
Decide what to do with
irrelevant information The researcher has to decide what to do with the irrelevant information. That is whether to ignore or use it to reconsider the codification procedures.
Coding the text The researcher has to decide whether to code the text manually, by reading the text and manually writing the occurring concepts, or by the use of software or both. Encoding with computer programs is of great help since it can examine large amounts of data in a wide variety of texts in a quick and efficient manner.
Finding Analysis When the coding is completed, the researcher has to decide what to do with the information from the texts that were not coded, whether to delete or skip such information or try to understand all information as relevant and important and uses them to reconsider, reassesses and perhaps change the encoding scheme. The researcher then prepares to conclude by investigating the data and make conclusions, and, if possible, make generalizations.
Source: Adapted from Rossi et al. (2014) However, there are practical and methodological difficulties with content analysis. There is often an enormous amount of data available that require a significant amount of time for analysis (Howell-Richardson and Mellar 1996; Gerbic and Stacey 2005). This challenge was overcome in this study with the use of computer-aided qualitative data analysis software (CAQDAS) called Nvivo 10. See Section 4.7.5, page 149 for details.
147