Conducting content analysis

The word "content" means the content (or content) of a document. Content analysis refers to a method for collecting quantitative data on a phenomenon or process under study that is contained in documents. A document is not only an official text (such as an instruction or a legal law), but everything that is written or spoken, everything that has become a communication. Books, newspaper or magazine articles, advertisements, television performances, film and video recordings, photographs, slogans, labels, drawings, other works of art, as well as official documents, of course, are subjected to content analysis.

The method of content analysis has been used by sociologists for over a hundred years. With its help, religious symbols and popular songs were studied, erotic films were distinguished from pornographic films, a measure of the effectiveness of political slogans (slogans used in election campaigns), advertisements and enemy propaganda was established, the characteristics of suicidal behavior that manifested themselves in suicide suicide attacks, stereotypes of consciousness social groups, identified the direction of demonstration of people of a certain nationality on television, the ideological background Newspaper articles, differences in interpretations of the same event in different media, explored many other topics. In recent decades, this sociological method has been borrowed and actively used by those representatives of the social sciences and humanities (lawyers, historians, journalists, linguists, literary scholars, political scientists, psychologists, economists, educators, social workers, cultural studies, library scientists, art historians, etc.) who are interested in establishing objective signs of a variety of human communications.

What documents are subject to content analysis? The answer to this question depends on the research program, the object, subject, purpose, objectives and hypotheses of the study. If, say, we have to find out the prospects for the strike of the workers of the enterprise (we continue this example), then the need for content analysis, at least for the minutes of workers ’meetings, decisions of relevant trade union committees, executive orders, laws governing strike strikes, etc., becomes obvious.

Content analysis does not eliminate the need for the usual (ie meaningful) analysis of documents. The first complements the second, their combination deepens the understanding of the meaning of any text. Content analysis allows you to find in the document that escapes from the superficial glance in its traditional study, but that has an important social meaning. The principal difference between these methods of analysis lies in the pronounced rigor, formalization, systematization of content analysis. It is aimed at developing a quantitative description of the semantic and symbolic content of the document, at fixing its objective features and counting them. According to a number of sociologists (Markoff, Shapiro, Weitman, etc.), content analysis could be called “textual coding”, since it involves obtaining quantitative information about the contents of a document based on its coding.

In addition, content analysis differs from all other ways of studying documents, in that it allows you to “fit” the content of a document into a social context, to comprehend it both as a manifestation and as an assessment of social life. “Writing” the document into the problem being studied involves identifying that: a) existed before it and was reflected in it, b) is present only in it, c) will be after it, i.e. will be the result of his perception by the addressee.

Formality, systematization and rigor of content analysis is manifested in the following. Before directly analyzing the text of the document, the researcher determines the categories of analysis, i.e. key concepts (semantic units), available in the text and corresponding to the definitions and their empirical indicators, which are fixed in the research program. It is desirable to avoid extremes. If too general (abstract) concepts are taken as categories of analysis, this will predetermine the superficiality of text analysis and will not allow one to delve deeper into its content. If the categories of analysis are extremely specific, there will be too many of them, which will lead not to the analysis of the text, but to its abbreviated repetition (outline). We need to find a middle ground and try to achieve that the categories of analysis are:

a) relevant , i.e. corresponded to the solution of research problems;

b) exhaustive , i.e. sufficiently reflect the meaning of the basic concepts of the study;

c) mutually exclusive (the same content should not fall into different categories in the same volume);

d) reliable , i.e. such that would not cause controversy between researchers about what should be attributed to one category or another in the process of analyzing the document.

After determining the system of categories of analysis, the corresponding unit of text analysis is selected. A unit of analysis can be taken: a) a word, b) a sentence, c) a topic, d) an idea, e) an author, e) a character, g) a social situation, h) a part of text combined by something that corresponds to the meaning of the category analysis. Sometimes, more precisely, when content analysis is the only method of information, one does not operate with one, but several units of analysis at once.

Then the account unit is set, i.e. quantitative measure of the unit of analysis, which allows to register the frequency (regularity) of the appearance of the sign of the category of analysis in the text. The counting units can be the number of specific words or their combinations, the number of lines, characters, pages, paragraphs, copyright sheets, text area, expressed in physical spatial values, and much more.

Conducting content analysis requires the preliminary development of a number of research tools. Of these, mandatory are:

1. content analysis classifier
2. the protocol of the results of the analysis, which has the second designation - the content analysis form,
3. registration card or coding matrix,
4. instruction to the investigator directly involved in the registration and coding of the units of the account,
5. catalog (list) of the analyzed documents.

The content analysis classifier is a general table, which summarizes all categories (and subcategories) of analysis and units of analysis. Its main purpose is to clearly state in what units each category used in the study is expressed. The classifier can be likened to a sociological questionnaire, where categories of analysis play the role of questions, and units of analysis answer. It is the main methodological document of content analysis, which predetermines the content of all the other tools of this method.

The protocol (form) of content analysis contains: first, information about the document (its author, time of publication, volume, etc.); secondly, the results of its analysis (the number of cases in which certain units of analysis are used in it and the conclusions that follow from this regarding the categories of analysis). Protocols are filled, as a rule, in coded form, but not for the sake of secrecy of the results of content analysis, but based on the desirability on one sheet of paper to fit all the information about the document so that it is more convenient to compare the results of analysis of different documents with each other. If a content analysis of a small number of documents is carried out in a study, then it is possible to do without coding and fill these protocols in an open and informative form.

A registration card is a coding matrix in which the number of units of an account is marked, which characterizes the units of analysis. The protocol of content analysis of each specific document is filled out on the basis of counting the data of all registration cards related to this document.

The formalization of content analysis and the complexity of its tools often repel novice sociologists from using this method. Thus, they impoverish their research, reduce its heuristics. If this method is used for the first time, numerous errors are made. Note the most common miscalculations.

1. Document analysis is ahead of the development of a research program.
2. The documents that are not related to the hypotheses of the study (having a similarity to the topic of the study only in name) are analyzed.
3. The authenticity of the document is not verified.
4. Not specified his authorship.
5. Its purpose is not fully taken into account.
6. The categories of analysis are not defined to the extent that allows to clearly distinguish the semantic units of the text of the document.
7. The categories of analysis are not subordinate and are not aligned with the definitions and operational terms that are recorded in the research program.
8. The categories of analysis are incomparable with the meaning and language of the text of the document being analyzed.
9. The units of analysis characterize the categories of analysis only externally, and not essentially, and therefore the units of analysis do not allow the content of the document to be identified in full accordance with the categories of analysis.
10. The analysis of the document is conducted without preliminary preparation of the whole complex of methodical tools.
11. The qualifier has shortcomings, compiled in violation of the rules of logic.
12. Registrars (coders) did not receive due methodical training.
13. Instructions for registration and coding is not complete enough, compiled by the researcher, who himself had not previously tested the toolkit.
14. The coding does not correspond to the program of mathematical processing of the research data.
15. The content analysis results are not cross-checked with information collected by other methods.

A source:

S.I. Grigoriev. Fundamentals of modern sociology. Ed. Altai State. un-that. 2001.

Continuing the theme:

Content analysis as a research method
Content analysis - method description
Content analysis: essence, tasks, procedures
Media Text Content Analysis
Advertising content analysis
Political Content Analysis