An exploration of tool support for categorical coding

In this paper we explore tool support for categorical coding of verbal and chat data. We consider tool support for manual coding, automatic coding by learning algorithms, and derive at a socio-technical approach in which human coders and learning algorithms together address the coding task. Given that a literature study suggests researchers devise, adapt and refine a wide variety of coding schemes, a categorization support system should handle and accommodate user defined coding schemes. Based on these ideas a prototype of the ChatIC tool was developed and evaluated with three coding schemes. For two coding schemes a sufficient inter-rater agreement between a human coder and the learning algorithms was reached.