As mentioned a few days ago, the past week saw the first round of the Open Government Dialogue, a three-phased e-participation initiative launched by the White House that aims to gather public input for the crafting of the Open Government Directive. From their May 21 announcement:
Today we are kicking off an unprecedented process for public engagement in policymaking on the White House website. In a sea change from conventional practice, we are not asking for comments on an already-finished set of draft recommendations, but are seeking fresh ideas from you early in the process of creating recommendations. We will carefully consider your comments, suggestions, and proposals.
Here’s how the public engagement process will work. It will take place in 3 phases: Brainstorming, Discussion, and Drafting.
Beginning today, we will have a brainstorming session for suggesting ideas for the open government recommendations. You can vote on suggested ideas or add your own.
Then on June 3rd, the most compelling ideas from the brainstorming will be fleshed out on a weblog in a discussion phase. On June 15th, we will invite you to use a wiki to draft recommendations in collaborative fashion.
These three phases will build upon one another and inform the crafting of recommendations on open government.
The first phase, Open Government Brainstorm, was convened by the National Academy of Public Administration and used IdeaScale, a crowdstorming or idea generation tool for large groups.
Based on my own Open Government Dialogue site activity tracking data from the past ten days, I did the following quick analysis:
1) Activity over time (incl. registered users)
Table: http://www.flickr.com/photos/planspark/3583067924/in/set-72157618585823580/
Graph: http://www.flickr.com/photos/planspark/3582188431/in/set-72157618585823580/
(Note that the “votes” curve uses a different scale in order to make it fit into the graph.)
2) Average user activity over time
Table: http://www.flickr.com/photos/planspark/3583071264/in/set-72157618585823580/
Graph: http://www.flickr.com/photos/planspark/3583198476/in/set-72157618585823580/
(Note that the “votes per user” curve uses a different scale in order to make it fit into the graph.)
What’s interesting is that up until 05/23 (two days into the initiative, at only several hundred registered users) average user activity was very high but dropped sharply over subsequent days as thousands of new — and much less active — users signed up.
For example, on May 23 at 8.32am (about 36 hours into the project), I measured the highest average activity per user:
- 2.5 ideas / user
- 3.0 comments / user
- 82 votes / user
As of today, May 31 at 12.08pm (almost a full 10 days into the project), average activity per user is much lower:
- 0.2 ideas / user
- 0.6 comments / user
- 11.7 votes / user
I see a real potential here how such user adoption and user activity information could be used in real-time to manage and optimize individual as well as overall participation levels, to distribute attention more evenly (e.g. away from the most highly-rated items) or to encourage collaboration among participants.
Related posts:
{ 2 comments… read them below or add one }
Note that there is probably a power law curve here. For example power user Robert David Steele Vivas posted far more than me — 8 Ideas, 150 Comments, 207 Votes (I only know this because I know him). What would the chart look like if the top 5-10 users were removed from the stats? What would the chart look like if it were based on median figures, rather than average?
And what would be the meaning of the differences between these charts, in terms of planning new sites?
Also, some participants have extensive networks to “get out the vote”, while others are just individuals. What is the meaning of that phenomenon for democratic brainstorming and rating?
I have seen sites that feature a random selection of ideas, rather than only most popular and most recent. This reduces the advantage for ideas that were posted early or by someone with a big network.
Thanks, Tom!
Unfortunately, I haven’t gotten hold of the user stats yet. Not sure if they are available via the IdeaScale API (haven’t seen them on the site). I agree that what we can expect is a power law distribution of user participation (with a fairly small group that is highly productive and a long tail of users with ever-decreasing participation levels).
Median figures would probably be more insightful but I haven’t had time to play with them yet.
I think the curve above goes hand in hand with a perceived (from my end, at least) increase in noise (off topic ideas or plain spam) that has hit the site since around the time the curves peaked. So maybe there’s some connection there that could be explored further.
I plan to add a rough measure of whether or not an idea was on topic and see how that played out over time (my guess: contributions were very focused initially but then strayed to all kinds of unrelated issues).
{ 3 trackbacks }