The ideal solution to analyzing a few hundred verbatim responses to a survey has yet to be found. Most text analytic programs aren’t much help here – once trained, they shine at processing tens of thousands of verbatims, but they are not much help with 200 verbatims. Yet 200 verbatims is enough to require hours of manual effort.
No wonder then that word clouds remain a ubiquitous visualization tool for verbatim comments. They hit a sweet spot of providing a convenient summary of pages of text without requiring extensive manual coding: they are like the arithmetic mean of verbatims. They tell you something about the data, but leave a lot out.
Revelation has a free tool that adds back in another layer of data. Unlike other free word-cloud tools, which focus solely on providing you attractive static images, Revelation WordTree makes word clouds interactive: click on the selected word, and see each occurrence of it, as separate branches from a tree. For instance, the following is from a public study we conducted of consumers’ impressions of different industries.
Clicking on the most commonly used word, people, shows a tree illustrating the word in context: under “people -> who ->” you can see “people who don’t feel like sorting out their taxes”, “people who need it”, “people who wouldn’t otherwise”.
It’s not just a way of visualizing the data. It’s a way of exploring the data. It helps you develop richer conclusions.
You can play with the first 300 or so verbatims from this particular data set to get a sense for how the tool works. Or “grow your own” word tree from one of your data sets at http://www.revelationglobal.com/wordtrees/.
Author Notes:
Jeffrey Henning
