Content analysis is a sociological method that involves reading a written text (document) in the language of mathematics. In general terms, we are talking about the translation of information contained in an ordinary letter into a statistical dimension. It is used for large amounts of information, for example, in the study of political programs of parties or candidates for deputies. Such program provisions are usually documents of a fairly large volume, therefore, in order to obtain the necessary information, they usually single out the subject of research, which is then “run” through the entire existing documentary array. To explain, consider a specific example.
What do we count?
So, we have party programs. We are interested in what ideological positions the participants in the electoral process take regarding the problem of integration, and how these positions differ from each other. We remember that content analysis is methodologically a section of statistics, like all practical sociology. The subject of the study was determined. Next, we need to understand what we will count. There are two options: either paragraphs where there are statements about integration, or statementswith similar content. For me, the last option is better, since there are many shades of statements, which means that it is not surprising to miss the hidden semantic loads. Figuratively speaking, content analysis does not like this: the slightest methodological error leads to biased results. The entire amount of work will have to be redone.
Statement classification
Now you should decide how the calculation will be made and for which groups. We break down party proposals into groups: the vector of integration (European-Eurasian); assessment of statements (positive - neutral - negative). It must be remembered here that content analysis (the example of the study demonstrates this) assumes the utmost accuracy, despite the subjective opinions of the sociologist himself. Therefore, you need to rely not on your own assessment, but on the prescribed context. It will be visible immediately. The following indicators follow: the number of utterances (for each group) and the total number of utterances. Then it is already possible to draw preliminary conclusions from the data obtained.
Counting procedure
Statements are grouped relative to each other and counted in such a way that their semantic and textual relationship is visible. For example, there are 100 statements, of which 90 are for Eurasian integration, but only 40 have a positive assessment. This means that a relative minority of parties are in favor of this vector, and (taking into account the median valueindicator) ideologically there is no certainty in this matter. This does not mean that the study was “mistaken”, content analysis is a fairly accurate method. The only question is that the concept of "integration" is associated not only with electoral sentiments, but also with other factors that need to be further explored.
Afterword
To avoid such a mistake, it is best to do a pilot, trial analysis. Then you can understand and clarify exactly what criteria will be used to calculate. The main thing is to clearly operationalize the concepts used in the analysis so as not to lose individual shades of statements. Content analysis is a painstaking work that requires special attention to the goals and objectives of the study. But, unlike the same mass surveys, it provides more objective results.
