I’m not sure what my guest blogs are going to look like – it’s hard for an academic to let the thoughts flow, so to speak, without thinking about what the reviewers will say…
As I said earlier, I’m excited about using the Cognitive Edge techniques in my research. As a first crack at it, we’re going to look at the very large database in the Aviation Safety Reporting System. If you don’t know what this is, check out http://asrs.arc.nasa.gov/ . Anyone involved in aviation - pilots, flight attendants, air traffic controllers, maintenance personnel – can file a report of an unsafe incident. The incentive for pilots is protection from penalties associated with their own actions (e.g., if they ‘bust’ an altitude, or do not follow an ATC directive). The incidents are de-identified and catalogued, and are searchable by keywords, dates, type of aircraft, etc. The reporter tells what happened in his/her own words. Seems like an idea story database to mine – but the authors are not available to answer any new questions on their reports. Any tips on indexing/signifying someone else’s stories after the fact are greatly appreciated.
Comments (5)
Kathleen, I do a lot of "after the fact" story-tagging now. It's certainly not as good as in-the-moment signifying.
It's easy to search for keywords and use those as tags. Anything repeated, or what Amazon calls "statistically improbable phrases" usually are good.
I heard in a speech from Dave S the idea of asking participants to title their stories, and use those titles as keys or signifiers.
Since the authors can't title them, you and your colleagues could come up with titles yourselves. A few alternate titles could add a lot of richness to the stories.
Regards, John
Posted by John Caddell | July 20, 2008 12:33 PM
Posted on July 20, 2008 12:33
Thanks, John - I will discuss titles at my next lab meeting.
Posted by Kathy Mosier | July 20, 2008 4:30 PM
Posted on July 20, 2008 16:30
I think Cynthia Kurtz has also done some work on generic story tags - ask Dave.
Also you (or better still the main users of the database) could also do an AVT extraction and use those as the tags?
Posted by David Williams![[TypeKey Profile Page]](http://www.cognitive-edge.com/blogs/guest/nav-commenters.gif)
|
July 20, 2008 9:30 PM
Posted on July 20, 2008 21:30
Hi Kathy,
This would definitely work and this is essentially the "Seeker" functionality of SenseMaker. If you want you can select a diverse set of incident stories that you would like to seek out perspectives on and then signify against a specifically designed signification framework. What I think you should consider doing is track perspective groups (i.e. jr. v. sr. pilots, different piloting histories, age, gender, etc.) and then see how interpretations of the select story set varies across perspective groups.
I think you would also want to focus on volumes of interpretations against exemplar incidents (i.e. many to one versus one to one which is what you get with signification at the source).
Drop me a note if you would like to bounce some ideas around further.
Posted by Michael | July 21, 2008 2:14 PM
Posted on July 21, 2008 14:14
Good ideas, Michael. I'd like to talk to you more as we get into the project.
thanx.
Posted by Kathleen Mosier | July 24, 2008 12:04 AM
Posted on July 24, 2008 00:04