Content Analysis in Political Research

This topic was the subject of the first Workshop on January 13, 2003. To read the outline for that talk, go HERE. Alternatively, you can use the menu below to guide yourself through topics relevant to content analysis in political research: A selected review of content analysis in political research: history and seminal studies Books about computer methods of content analysis Sources of political texts in digital form Computer programs for content analysis Selected extended applications of content analysis to political texts

Rationale Inevitably, research on political topics--like research in any field- involves acquiring and cultivating methodological skills. But so many research methods exist that no graduate program can be expected to offer formal instruction in all of them. Faculty will differ on which methods deserve to be taught, and some methods--such as content analysis--are rarely taught in any political science department. Nevertheless, the systematic analysis of documents or speeches is often used in political research--but not as much as it deserves. Although a "google.com" search of the internet for "content analysis" and "political science" turned up 5,610 hits, a similar search for "content analysis" and "sociology" the same day ( (on January 6) yielded 12,100 hits. It seems that sociology (a smaller discipline in numbers of faculty and students) relies on content analysis more than twice as much as does political science. Given that so many political documents are a matter of public record, I think that researchers should know more about content analysis methodology.

In the second edition (forthcoming) of his 1980 text, Content Analysis, Klaus Krippendorff defines content analysis as "a research technique for making replicable and valid inferences from texts (or data) to a context of their use." That is, it refers to methods for inferring meaning from the text. To narrow the workshop's focus, I will concentrate on computer methods for analyzing textual materials, such as presidential speeches, government documents, party platforms, newspaper editorials, professional publications, and so on. By analyzing words rather than numbers, we will, in effect, be using computers in a form of "qualitative analysis." This term helps distinguish comptuerized content analysis from "quantitative analysis" of numerical data. Qualitative analysis of textual materials can take two different routes: information retrieval or content analysis. Information retrieval gives a modern label to the traditional work of scholars: gathering, indexing, storing, and accessing information from various sources. Every scholar needs to consult sources and retrieve information in the course of research. Computers can help a great deal in these activities. In fact, computer techniques for handling source material often open up entirely new research opportunities beyond taking notes and managing bibliographies. Content analysis involves systematic study of messages conveyed in natural language text. In political research, it usually results in straightforward tabulation of words and phrases used to communicate between political actors. Often these tabulations are subject to elementary statistical analysis. More sophisticated applications involve evaluating both the content of the message and its style of presentation.