Lexicoder performs simple deductive content analyses of any kind of text, in almost any language. All that is required is the text itself, and a dictionary. Our own work initially focused on the analysis of newspaper stories during election campaigns, and both television and newspaper stories about public policy issues. The software can deal with almost any text, however, and lots of it. Our own databases typically include up to 100,000 news stories. Lexicoder processes these data, even with a relatively complicated coding dictionary, in about fifteen minutes.

The software has, we hope, a wide range of applications in the social sciences. It is not the only software that conducts content analysis, of course – there are many packages out there, some of which are much more sophisticated than this one. The advantage to Lexicoder, however, is that it can run on any computer with a recent version of Java (PC or Mac), it is very simple to use, it can deal with huge bodies of data, it can be called from R as well as from the Command Line, and its free.

The Lexicoder Sentiment Dictionary is designed to be used within Lexicoder, but can be used in any content-analytic software package. Indeed, the dictionary is now included directly in the quanteda package for R, build by Ken Benoit The dictionary is discussed and tested in detail in Young and Soroka 2012.

Our work was greatly aided by conversations with several others involved in automated content analysis. We are particularly thankful to Will Lowe and to Ken Benoit.,