Narrative analysis with benchmarking

I just added this cool feature to our free, online story analysis tool that lets you build two story collections, then use one as the benchmark for the other.

Search.

You can use the story text or any of the associated survey questions.

Click to analyze.

Review each collection side by side, before merging. In this example, I’ve searched for stories that contain “lover” to compare against the larger reference group of stories that contain the word “wife” or “husband”:

I was surprised to find that stories about “lover,” “wife,” or “husband” are all more negative than our typical story in the collection of 57,000 stories from East Africa (so the icons appear red). The default benchmark (all stories) isn’t very helpful. But watch…

Merge.

Click “Merge A/B icons” and it will divide the left “A” side by the right “B” side and re-analyze instantly.

Whoa – something surprising happens when you use the wife/husband stories as the reference collection to compare ‘lover’ stories from East Africa: Stories about lovers are more positive and much likely to be told by men. This is quick-n-dirty gender analysis in a jar, using stories, frequencies, and meta data.

Just because you’re interested, “wife” vs “husband” is also revealing: Men speak more negatively about their wives than wives do about their husbands. And with some more sophisticated filtering, you can verify this precisely.

Go play. The tool is online and takes just a few minutes to learn:

http://djotjog.com/c/compare

Next: A longer explanation with many examples