Outputs from Challenge 2: Establishing trust & credibility

[Chris van der Walt]

Hi everyone,

Here is the second set of outputs from the good people at Adaptive Path focusing on Challenge 2: Establishing trust & credibility.

Lot of good stuff here with a focus on on establishing reputation and providing mechanisms to surface user interactions in the system as a way of establishing credibility. In our discussion yesterday with the team some other issues were also raised:

- How much personal information is absolutely necessary to share in order to establish trust among others?
- Vouching for someone should mean something. A way to establish value is to limit the number of vouches someone can give. Perhaps on an annual basis?
- Trust and verification are very different, needs to be explicitly indicated as such
- Trusting a group is different than trusting an individual.
- Who owns the data for each instance of HunchWorks? How can people trust that it will not be used without their knowledge of against them? Should instances be verified by a central organization?
- How can trust be "imported" from other networks? Having 100,000 Twitter followers could establish a level of validity among certain people.
- There likely needs to be a range of verification rather than a binary demarcation. Verification will also be subjective.
- A hunch's credibility score may ultimately be its verification score as well.
- EVE Online has a very interesting model for educating new players in the system, its game mechanics and culture. There are whole guilds that are dedicated to teaching new players how to succeed.

We are delaying the next design sprint for a week due to work commitments. In the meantime please give us feedback!

Cheers
Chris

[Jen Ziemke]

Dear Global Pulse Team,

Wonderful work on these outputs.

My concern at this point is: the team has done much of the work outlining how "trust" and "verification" works in the system and can be created and enhanced.

Now it seems to me you need to flesh out in the same sort of way how to "lower the bar" enough for people to actually use the system and post hunches that they themselves aren't 100% certain are correct. In other words, there is a tension between wanting to make onesself "trustworthy" by only posting completely accurate (and likely banal) hunches on the one hand, but on the other hand the whole premise of the system is to get people to share their unverified hunch, just their sense about something, that could then be later falsified. If people are afraid that their falsified hunch would lower their crediblity score, they may never put forward a "crazy" hunch idea, even though it may be precisely the kind of information you want in the system. I think you want to build a culture of "even if your hunches turn out not to be quite correct, that is ok, we welcome your participation." So, while GP needs to be concerned about the trust/credibility issue, I think you need to be equally concerned about encouraging partcipation in the system in the first place. In other words, who would creatively brainstorm something if they felt they were being judged, assessed, etc? My belief is that the key is to build something that seems easy to use, that generates a warm and welcoming atmostphere, & is not an overly intimidating space based on building personal reputations.

Jen

--

Jen Ziemke, Ph.D.

Co-Founder & Co-Director
http://www.CrisisMappers.net

[Chris van der Walt]

Hi Jen
I agree that we need to make sure that there is no perceived barrier to posting unusual or crazy hunches. However, I don't think that the credibility and reputation mechanisms suggested are likely to do that. There is nothing to stop anyone posting hunches that are unusual or poorly backed up with evidence initially - it is the purpose of the system to help those hunch providers to identify other users who can help to verify or refute the hunches. Rather we want to provide metrics that will help to identify helpful users with complementary skills easily and the reputation systems will be focused on positive metrics. We certainly don't want to give the impression that the space is a judgmental one. I am hoping that as we move to the particular hunch mechanisms themselves, some of this might be clarified.
Thanks for flagging this as an issue to watch out for - we will feed it back into the design process.
Cheers
Chris

[PJ Onori]

Very good point Jen. Here are my thoughts on this issue.

From my perspective, we're walking a very thin line. We want people to feel comfortable throwing out ideas that may be seen as out of left field, but from my point of view, we also want to keep the signal to noise ratio fairly high. Even if a hunch may be a seen as a little crazy, it also needs to be thoughtful and based on some level of consideration. I agree, a person's reputation should not necessarily get dinged for being wrong, but perhaps it should if it's simply creating confusion.

I would also suggest that in creating a system of trust we could actually increase how comfortable people will be to throw out crazy ideas. If mutual trust exists, people are much more likely to both go out on a limb with each other and share something they may not think is fully formed. It's through fostering trust that people can be comfortable to creatively brainstorm issues. I could also see this type of participation occurring quite frequently amongst smaller trusted networks depending on the users' personalities.

All that said, there is no intention to create a system where people are marked as "good" or "bad". It's something that will need to be handled sensitively and respectfully.

Thanks again!

[Sara Farmer]

Or perhaps we could weight reputation more on evidence than on hunches.  That way, people would be more free to post speculative hunches, but would be expected to add sensible evidence to them. Sj. --- On Sun, 5/6/11, Chris van der Walt <vande...@un.org> wrote:

[P.J. Onori]

That's a great idea Sara.