DiscoverText

DiscoverText is an updated version of the extensively popular Coding Analysis Toolkit (CAT), allowing open-ended textual data

Starting to use DiscoverText

Installing the software

DiscoverText is used through a web interface. This is a software as a service with a free trial, after which there is a monthly fee.

Setting up a project

A project is a collection of documents, each of which is coded. One must create a project before data can be imported.

Importing documents.

DiscoverText allows importing several documents (as text files), or a spreadsheet collection of data. These collections are known as Data Archives, from which you can created Buckets and Data sets which are subsets of filtered data. Data sets are needed in the coding stage.

Doing data analysis

Create codes

in DiscoverText codes must be predefined. These are known as code sets. A code set is a list of codes, each having a specific analytical code name, a code description to define what the code means and a short-hand key which allows the coder to use the code quickly via the keyboard. The coder is only able to use the codes defined on a code set.

Coding data

The coding interface allows the coder to quickly code full coding unit with a given code from the code set. It is also possible to segment the unit with a specific code, i.e., conduct coding at the level of individual phrases or sentences. This is done by selecting the segment requiring annotation and then giving the specific text (called trigger text) a code from the code book.

Analysing reliability

When conducting coding from theory-driven lists, it is important to evaluate how well coders agreed with each others. There are many ways to evaluate this inter-rater reliability, including Cohen's kappa. DiscoverText allows you to automatically calculate such evaluations through validating datasets.

Automating content classification

DiscoverText provides support for automating content classification work. We will cover some aspects of it in the automation module.