Schwablearning.org 's theory of change: we want to allow for the life-long success of a child with learning or attention disabilities We wanted to create change through:
â€¢ Quality, research-based information
â€¢ Reduce feelings of isolation
â€¢ Change attitudes about success with a learning disability
â€¢ Taking action
We measured this through monthly anecdotal reports, a one-year survey of members, (and a third way that your liveblogger sadly missed.) One important measure ended up being that when their community sites shut down other sites formed, based on communities that the site had fosteredâ€”Scott wished they'd known about those relationships prior to the end of the lifecycle.
Four points for measuring impact:
1. When thinking about impact: define it. Socratic way, break it down, define it, easier to measure
2. Measure boldly. Don't be afraid to ask " Can we measure thi?" Use indirect methods, measure something related. Brea it down, measure the pieces. Don't be afraid of having perfect science. Reduce your level of uncertainty about it. Get closer to the answer, rather than thinking you can't answer it.
3. Ask for help.. Look outside. Some one might have already measured it. Conf: Communities and Technology. Other people have already solved your problems.
4. Fail informatively. (Clay Shirkey) Measure what you're doing, and if it doesn't work, share where you failed (and succeeded)
Q (DoSomething): Challenge of multiple audiences for your research. How do you deal? A: Understand the impact form the asker's perspective, ask the questions that way, share the data that way.
Q (Isaac/ SquarePeg): We're developing a tool, talking to orgs, they're pulled in lots of directions by different tools. Isaac tells orgs to lay down their theory of change, then pick your tool.
Scott suggests resources for measurement:
Measuring specific impact: Google it
Susan Colby SSRI article "Zeroing in on Impact"
Douglas W. Hubbard How to Measure Anything
KD Payne, metric queen
Q: How to measure offline impacts?
A: If you lack tech to ID individual speakers, figure out a way to do that, say qualitatively. With 3-7K posts per mo, mine them all with keyword hitlist. Example: Scott tracked "alone" " not alone" - and with this data started to measure how they were reducing isolation. If your sense of impact, for example, is the quality of conversation, then look for word patterns, filter all the conversations, pull out the names, see clusters and patterns about who is having those types of conversations, look at who listens to them, see how they are impacted.
(and back to Rachel's voice)At this point, we're into a pretty conversational Q&A, so I'm going to post this now...typos and all. Signing off, this is Rachel livebloggily yours!
[Added later] Great tip from a participant: Track Qs that come in through widgets, if that question results in a story, then report back to the site that hosted the widget. CLose the loop!