Social Capital and Trust

Establishing Trust

I, like most people it seems, had a little trouble with some of the readings. Some of them were full of statistical measurements and analysis, which I don’t have a very good background in.  I found “A Survey of Trust Use and Modeling in Real Online Systems” by Paulo Massa to be the easiest to read even if it was a bit lengthy. The paper surveys different ways websites establish trust between users.  Establishing trust is especially important for places like eBay, where goods are auctioned off for services. You want to know if the seller is legit and if you are going to get the item you want.  However, trust is used in many other websites as well. uses trust to establish how trustworthy a reviewer of a product is.  LinkedIn establishes trust by having others in your network recommend you.  The paper also mentions P2P websites, which I would not have considered before reading the paper.  It makes sense though, as you would still want users who provide trustworthy content and also remain connected to the network for downloading.  I also had a brief introduction into PGP in ICS 623 (Data Security), but did not hear about the web of trust.  The idea of verifying identity through the web of trust is quite novel and almost seems ahead of its time (according to Wikipedia, the idea was initially proposed by Phil Zimmerman in 1992 for PGP 2.0).  It’s very analogous to how we find friends in SNSs today, where we find people with similar interests through our initial network of friends.

Digg vs. Reddit

I’ve been a user of Digg for quite a while.  I’m not as involved with the community as I’d like, but I do rate stories that interest me as well as the occasional funny comment. I’ve been around enough to know of the Digg vs. Reddit rivalry that has been going on for a while.  They’re both very similar sites; they both are social news websites that display news stories submitted by the community and they have the community decide what news makes it to the front page.  In a way, I guess they’re both similar to Answerbag from our previous assignment since users submit questions to the community and have them rated and posted to the front page.  Anyway, Digg has become the much more mainstream product (Kevin Rose, one of the creators of Digg, has become a rock star as of late) while Reddit has remained fairly niche.  Thus, for the first website, I chose to take a closer look at Reddit to see how it compares with Digg.

Creating an account on Reddit was very simple. No need to even submit an email address; just a user name and password (along with a CAPTCHA) and your account is created. The stories on the front page can be rated up or down by the community.  It’s very much a news site from the Massa paper, but unlike Slashdot, the moderators are other members of the community.  Like Digg, links are voted to the front page after they reach some threshold. They also use the word “karma” to indicate how involved a user is, which might be something they borrowed from Slashdot.  There are two types of karma, regular karma and comment karma.  Presumably, karma is the sum of the scores of all links the user has submitted and comment karma is the sum of scores of all comments a user has.  It’s interesting how the two are split apart. I would think that you want to friend people with high regular karma, since they seem to submit all of the hot stories.  You also don’t see if the user has received any negative feedback, since you just see a raw positive score.  I think that negative feedback for users in Reddit could be highlighted so that users can identify people who only gained social capital in order to spam the users of the community.



Overall, Reddit is an interesting site that I think I might be visiting a little more.  Digg has become a little more mainstream than I’d like.  Seeing programming questions like this on the front page of Reddit reminded me of the stories that used to make it to the front page of Digg.


There are other ways of sharing cool stuff you find on the web with others.  I have heard a lot about StumbleUpon, but have never used it.  During the Manoa Experience a few months ago, a high school student mentioned something that they found through StumbleUpon.  StumbleUpon is a website where you randomly visit (or stumble upon) sites rated by the community.  Using their toolbar (they have a bookmarklet and a Firefox extension as well), you press stumble to visit a new site. Once there, you can give the site a thumbs up and/or write a review of the site.  You can also view reviews from other users of the site to see what they thought.  There, you can meet friends who may or may not have similar interests to you.  There, you can give them a thumbs up or subscribe to their links.


It seems that both StumbleUpon and Reddit are similar in that there’s an algorithm tailoring the content to your tastes.  StumbleUpon is neat in that you can just use the toolbar and not have to visit the main page unless you want to accept friend requests and the like.  In Reddit you still peruse the headlines, while in StumbleUpon you just click the “Stumble” button to move through recommended sites.  StumbleUpon seems to use a 5 star system in their toolbar, even though they mostly use a thumbs up/down system in the site.  The toolbar could show the ratio between positive and negative reviews as well as the star rating.

Using Geographical Information to Establish Trust

While the readings were somewhat difficult to get through and understand, they did help me decide on my final project.  I’ve been doing a lot of mobile phone programming (specifically iPhone) and want to investigate mobile social software.  In the LILT lab, we already have an iPhone application called iGeoRSS (iTunes link) in the app store.  iGeoRSS is an app that acts as a GeoRSS reader, which is just an RSS feed with location information.  Using the application, you can see how far each news item is from your current location.  We also put in a commenting framework so that you can comment about the news item you’re viewing.  The comments are also geo-tagged so that you can see the location of the commenter.  Currently, the comments are ordered by their distance from your location, although they could also be ordered by distance from the location of the news story.  The idea is that you might trust the opinions of people who are closer to the news event than those who may be farther away.  On the other hand, you might be more interested in the comments of others in your area than you are of people farther away.  It’s all anonymous right now, but having users who comment and gain social capital is something I’m interested in looking into.


4 Responses to “Social Capital and Trust”

  1. 1 Michael Claveria March 7, 2009 at 1:26 pm

    I wrote a 3 paragraph response only to have it lost when there was a posting error. Anyway, I’m very pissed off so I’ll try once again. In stumbleupon I noticed that the bottom of the page displays users who are currently logged on. All of those users are from Hawaii. Are people expected to built rapport with others from Hawaii? I clicked on some profiles and I liked how the different tags of interest were of different sizes. However, I didn’t get a whole lot of information about a user from the profile other than a picture, a few words and uploaded photos, videos and favorites. I see some features that support bridging such as the stumbling and getting to know other people, but there’s not a lot for bonding social capital.

    I didn’t care for the web design of reddit as the page seemed very cluttered. I’m not sure what to make of Karma. If I clicked on a user and saw his karma was 302, what does that mean? I’m curious as to how karma is calculated.

    Your final project idea sounds promising. Are you going to study any sites as well? How can one measure whether trust is stronger between people who are in geographically similar areas?

  2. 2 karhai March 8, 2009 at 12:35 am

    I have a similar question to what Michael already asked — how are you measuring trust? Or more specifically, what constitutes trust when a user is reading a news story? Whether or not he believes the story is true? Trust as a form of social capital is required for any news site to be successful, but I’d argue that any user who has setup his RSS feed to receive news from a particular site must have a certain level of trust to have bothered to do so in the first place.

    I do like that StumbleUpon and Reddit tied right into your idea for a final project. They do seem to follow a similar path as Answerbag, where user generated content is rated by popular vote. I wonder if you could do your geo-focus study using data from sites like StumbleUpon or Reddit… do their news submissions have any kind of geo-coding? Could you see if responders or voters have any geo-bias in the news stories they rate?

  3. 3 keokilee March 8, 2009 at 11:00 am

    Reddit is a little cluttered, but you can personalize your feeds. Haven’t messed around with it too much, since I usually just read things out of my RSS feeds. I don’t think that StumbleUpon or Reddit use any sort of geo-coding. I have seen other news sites that do some geo-coding, so a review of sites like that needs to be done as well.

    My original project idea was to survey mobile social software and see what makes them so compelling. I think that with this project, I’ll still do that with an eye toward the social capital/trust mechanisms that the sites employ. As for trust between commenters, we started off by saying that the person who comments closer to the news story is perhaps the more reliable one. It’s a simple and straightforward approach, but the user doesn’t gain much social capital from doing so. I’m thinking of moving to a model where we highlight top commenters in regions by measuring the amount of comments they made and their distance from the original stories. We could also throw in a comment rating system like Digg or Reddit too. I’m not sure if I’ll be able to actually implement all of this by the end of the semester, but I’d like to get the research and groundwork done and maybe implement it over the summer. As far as the stories, Karhai does bring up the point that there already is some level of trust for someone to add a RSS feed. Another thing that would be interesting is to create our own GeoRSS feeds highlighting the most commented stories in regions around the world. That would create a trustworthy source for users that shows what’s going on in their area.

  4. 4 chakroff March 9, 2009 at 1:03 am

    It sounds like you’re taking an original approach to the project, and it would be interesting to watch the trust mechanisms go into effect. I’m curious to see if users prefer trusting those closest to the news over those closest to the user. I can also imagine that it would be popular to get the perspective of someone from another culture (I quite vastly prefer BBC World News to NBC Nightly News, especially since Tom Brokaw retired.), and it would be interesting to see how those users both far away from the story and from the reader earned trust. For example, I don’t live in Michigan now, but I’d likely trust me over Jenna if we were asked about plant closures in the states. (…actually, Jenna might be a bad example, I think they’re closing plants in Germany, too, so the whole thing might be on her radar already.) dude, the point is, would there be a way to rate trustworthiness on specific issues? Whatever angle of attack to take, good luck!

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

Recent Comments

RSS Other posts in ICS 691 Social Computing

  • An error has occurred; the feed is probably down. Try again later.

%d bloggers like this: