WordPress database error: [Table 'i203.is203_users' doesn't exist]
SELECT * FROM is203_users WHERE ID = '1' LIMIT 1
Week 12
Apr. 10th: Reputation in Online Settings
Kollock, P. 1999. “The Production of Trust in Online Markets.” in Advances in Group Processes, vol. Vol. 16, edited by E. J. Lawler, M. W. Macy, S. R. Thye, and H. A. Walker. Greenwich, CT: JAI Press.
Resnick, P., Richard Zeckhauser, Eric Friedman, and Ko Kuwabara. 2000. “Reputation Systems.” Communications of the ACM 43:45-48. ![[PDF]](http://dream.sims.berkeley.edu/~jantin/is203/wp-content/acrobat_pdf_icon_sm.gif){kind=icon}
Chapter 5 in Rheingold, Howard. 2002. Smart Mobs: The Next Social Revolution: Perseus Books.
Apr. 12th: Social Identity in Online Settings
Donath, Judith. 1998. “Identity and deception in the virtual community.” in Communities in Cyberspace, edited by P. Kollock and M. Smith. London: Routledge.
Burkhalter, Byron. 1999. “Reading Race Online: Discovering Racial Identity in Usenet Discussions.” Chapter 3 in Communities in Cyberspace, Kollock and Smith, eds.
Chapters 1-2 in Turkle, Sherry. 1995. Life on the Screen. New York, NY: Touchstone.
O’Brien, Jodi. 1999. “Writing in the Body: Gender (Re)Production in Cyber Interactions.” Chapter 4 in Communities in Cyberspace, Kollock and Smith, eds.
January 2nd, 2007posted by:
WordPress database error: [Table 'i203.is203_users' doesn't exist]
SELECT * FROM is203_users WHERE ID = '1' LIMIT 1
8 Comments Add your own
1. yliu | April 9th, 2007 at 7:01 pm
Reputation is a fundamental heuristic we all use to make judgments about people with whom we interact with. “Past behavior is indicative of future behavior” or some variant thereof gets thrown around a bit. A heuristic, to be sure, but generally a useful one. The Resnick and Rheingold papers pick up an interesting point however. To be truly effective, reputation systems require a lot of participation and a lot of data. A sparsity of feedback, much like newcomers to eBay, makes heuristic assessment of trustworthiness difficult. Much of this is focused on making feedback effortless and costless, and reputation systems easy to use. On the flip side, it seems that unification of reputation data from across information systems is another approach to the problem. By doing so, we even gain additional dimensions for which a person can be assessed on (”so he’s a great eBay buyer of various trinkets, but can I trust him to ship me a rare book…” etc).
In the past, I’ve always thought of this from a privacy perspective. Unification of data always scares people, for various reasons. Further, some people just don’t want to share this information (and the system should accordingly tolerate but punish non-participators in some way…it wouldn’t do to have too many free-riders in the reputation system, for fear of the entire thing collapsing). I didn’t know that eBay has claimed these ratings as proprietary information for itself. I find it interesting that eBay considers my reputation its property.
The ubicomp reputation machine/tracker comes up again. I’m still intrigued. What kind of interface would be required for people to be able to rate others in the real-world and in real-time? Being able to rate others after interacting with them in some exchange, or being able to see if I can in fact trust the guy at the airport to watch over my luggage… I think there’s an interesting use case there. Can we deploy this as an application on a cell phone? A special Wifi/Bluetooth equipped reputation device? What privacy implications will there be? How can this be managed, if at all?
The discussion of Usenet traders ring an bell. Bootleg exchanges have given some way to P2P software. P2P exchanges in particular is interesting because it abstracts away some of the reputation-based trading problem from Usenet, but only in so far as to place them in the machine/agent world. BitTorrent clients are still designed with things that resemble reputation systems, prioritizing those clients that are actively sharing. They use checksum mechanisms to ensure data integrity and automatically ban defectors/betrayers that hand them bad data. There are blacklists and blacklist management plug-ins (to fend off suspicious RIAA spiders, apparently). In many BT trackers, there are upload/download ratios kept on each agent, shutting down those agents that have very low ratios as “untrustworthy” free-riders. Of course, reputation of “agents” here are only reflective of their operators, the humans behind the P2P software. But it seems an interesting precursor to reputation systems that function for agents as well as humans.
At the same time, it seems that sites like Craigslist have taken on much of those newsgroup functions in other areas. I think it is telling that Craigslist strongly recommends “face-to-face, local transactions” for any exchanges made there. eBay and co have become more or less global marketplaces, thanks to no small part, I think, to their reputation tracking systems.
April 9th, 2007 23af at 7:01 pm |WordPress database error: [Table 'i203.is203_users' doesn't exist]
SELECT * FROM is203_users WHERE ID = '1' LIMIT 1WordPress database error: [Table 'i203.is203_users' doesn't exist]
SELECT * FROM is203_users WHERE ID = '1' LIMIT 12. Ken-ichi | April 9th, 2007 at 7:53 pm
On Resnick et al.
Given that Coye has mentioned several times that reputation systems generally don’t work, I was surprised to find this article touting the surprising functionality of online reputation systems, particularly the one used at eBay. That is, despite the many opportunities for fraud in anonymous open systems like eBay, the amount of fraud that actually occurs is far less than one might think. I found it interesting that reputation systems shouldn’t work from a theoretical perspective, but is this because these theories are predicated on everyone being jerks? The authors mention problems with reputations systems like the difficulties of eliciting negative ratings, and the difficulty or impossibility of porting reputation across separate systems, or separate online identities, but clearly the majority of users get by without the more fine-grained ratings that might be had from more abundant negative rating data, and although some people may manipulate the system with multiple identities, there must not be enough of them to make the experience too risky. What kind of assumptions about the proportion of fraudulent users do theoretical models hold?
April 9th, 2007 at 7:53 pm |WordPress database error: [Table 'i203.is203_users' doesn't exist]
SELECT * FROM is203_users WHERE ID = '1' LIMIT 1WordPress database error: [Table 'i203.is203_users' doesn't exist]
SELECT * FROM is203_users WHERE ID = '1' LIMIT 13. daniela | April 10th, 2007 at 11:41 am
I also thought Resnick’s claim was interesting that in order to design working reputation systems one should design for the worst-case scenario – usage by a cheat. I don’t remember Coye arguing against reputation systems. There are clearly ways these systems can invite contributors and discourage freeriders, allowing for more successful interactions. It’s a shame that Amazon was limiting the portability of their system, shutting down the rating-import system with eBay. It seems that more centralized systems would allow for even more successful interactions since they account for more use and a longer time frame with which to build “trust”-like relationships.
I was a little surprised by Rheingold’s characterization of the e-Opinions reputation system. The site has likely changed since he wrote the chapter in 2000, but currently the most notable mark of reputation is the “Top 100”, “Top 50” or “Top 10” marks associated with certain rators. These marks are completely determined by the top-down structure – a moderator working for e-Opinions decides, not a user. But the user typically never knows. This brought up an important point in Rheingold’s chapter – does it matter how transparent the process is that determines each person or post’s reputation? How do we know to trust the moderator? There must be a difference between community driven marks and marks determined by a moderator. Yet I doubt many people using e-Opinions need to know where those marks come from as long as they trustworthy and worth reading over time.
April 10th, 2007 at 11:41 am |WordPress database error: [Table 'i203.is203_users' doesn't exist]
SELECT * FROM is203_users WHERE ID = '1' LIMIT 1WordPress database error: [Table 'i203.is203_users' doesn't exist]
SELECT * FROM is203_users WHERE ID = '1' LIMIT 14. karenhsu | April 11th, 2007 at 1:34 pm
In response to Resnick’s “Reputation Systems” paper, I’d like to add that PayPal not only provides a complementary service that facilitates the purchasing of items won on eBay, but it also provides somewhat of a safety net for evil transactions. The PayPal Buyer Protection policies provide its users up to $2,000.00 USD of purchase protection on items that were (1) never received or (2) significantly different from what was described in the listing. I’m sure this has contributed to buyer confidence and, with more transactions, has also added to eBay’s reputation system success.
Something interesting to think about is the integration between transaction history/feedback and information gathered from social networking websites, such as Facebook. A possible user conversation with oneself might ensue like, “Oh, well.. this person is a 4th degree friend of mine, I might be able to trust doing business with her,” or “Wow, this person is in a lot of shady groups, he might not be so reliable,” etc.
April 11th, 2007 at 1:34 pm |WordPress database error: [Table 'i203.is203_users' doesn't exist]
SELECT * FROM is203_users WHERE ID = '1' LIMIT 1WordPress database error: [Table 'i203.is203_users' doesn't exist]
SELECT * FROM is203_users WHERE ID = '1' LIMIT 15. elisa | April 13th, 2007 at 10:43 pm
A wealth of great readings this week, very stimulating and by very good writers. I particularly liked Rheingold and Donath. The “distributed reputation systems for ad hoc wearable computer communities” immediately reminded me of Cory Doctorow’s science fiction novel Down and Out in the Magic Kingdom. The world he depicts is a world that has defeated death and poverty, and where people instead of cash/jobs etc. have Whuffies. Whuffies are a sort of reputation currency that everybody wears at all times. When one does something good, worthwhile, positive, other people send him whuffie points, that are like approval rates; when one misbehaves, he gets whuffie points taken away. When people meet, they instantly check the other person’s whuffie’s rate and behave accordingly. Because in this world there’s no more money, whuffies are the currency – if one has plenty, one gets the best tables in restaurants, nice houses, etc, if not, one is condemned to a life of misery and being shunned. Talk of the Shadow of the Future! But it doesn’t end here. A fan of the book decided that Whuffies (incidentally, a truly horrendous name) were a really good idea, and created them for real. So now there’s an open source something “creating a marketplace for trading and rewarding favors for your friends and like-minded strangers.” Very bizarre, mostly because the impression I had from the reading the book was similar to a 1984 world dominated by bigots, not a world where trust and friendship prevailed and good behavior was rewarded. I guess that my trust level is low, on- and off-line. Back to Rheingold, I found the Ultimate Game experiment very interesting, with its findings about the cost of a reputation for being a sucker, especially morally, but since I am not particularly articulated, I won’t expand on this point.
Donath’s article was also very rich and alluring. Her writing and choice of examples made it extremely readable, while conveying profound concepts. I am intrigued by the contrast between the inherent unity of the physical self, anchored by the body, and the “multiple personae sharing a single progenitor”. I wonder how people keep all these personalities straight, since most of us don’t even seem to keep their usernames in order…
2020 April 13th, 2007 at 10:43 pm |WordPress database error: [Table 'i203.is203_users' doesn't exist]
SELECT * FROM is203_users WHERE ID = '1' LIMIT 1WordPress database error: [Table 'i203.is203_users' doesn't exist]
SELECT * FROM is203_users WHERE ID = '1' LIMIT 16. jilblu | April 15th, 2007 at 9:09 pm
On Rheingold and reputation:
Rheingold’s piece was interesting to me because it discussed reputation in knowledge-sharing systems that were not related to e-commerce or the exchange of goods. Some examples include Epinions and Slashdot. In these examples, since users were looking for information instead of purchasing goods, they did not run the risk paying too much for low-quality goods. Instead, they ran the risk of being misinformed, or being deluged by unhelpful information. In knowledge-sharing systems, reputation is used to prevent low-quality information from overwhelming the system, resulting in a “tragedy of the commons.”
I used to be a devotee of Chowhound.com, a site for people obsessed with the minutiae of finding the best things to eat. Although, the site does not have an explicit reputation system for its contributors, I found that I quickly “got to know” many contributors by reading their posts, and figured out who’s posts were most worth reading. Factors I considered were quality of posts and frequency of posting. One contributor, Melanie Wong, was especially prolific; she commented on a large proportion of all posts. With her posts, she often included a photo of the dishes she had eaten; this was a few years ago – before the prevalence of digital cameras and camera phones – so everyone was impressed with her diligence.
Because the site did not have a built-in reputation system, I felt that “getting to know” the contributors on Chowhound was similar to getting to know a new group of people. Rather than being able to see immediately what most people thought of a particular contributor, this was knowledge that was built up over time - by reading many, many posts, by reading in between the lines of a conversation, and by getting to know someone’s writing style.
Back then, it didn’t occur to me to wonder whether or not people contributed using different identities, and I’m not sure I would have cared. The risk I was taking was pretty small – the pleasure of reading about an obscure food item or a surprising restaurant find far outweighed the risk of its being an inaccurate reporting.
April 15th, 2007 at 9:09 pm |WordPress database error: [Table 'i203.is203_users' doesn't exist]
SELECT * FROM is203_users WHERE ID = '1' LIMIT 1WordPress database error: [Table 'i203.is203_users' doesn't exist]
SELECT * FROM is203_users WHERE ID = '1' LIMIT 17. bindiya | April 15th, 2007 at 10:58 pm
It was extremely interesting to read Resnick’s paper. I was amused to read his ending comment which defined reputation systems as “theoretically flawed and practically challenged system that nonetheless appears to perform miracles.” I think is important that its is specified for which dimension the person/object is being rated for, since being rated for x dimension is not at all related to how a person should be associated at dimension y. Resnick gave a good overview of the problems and challenges related to reputation systems. It is important aspect to keep in mind that it should not be very difficult for the user to provide feedback on reputation systems. If it’s not easy to provide ratings, most people are not going bother, and this would bias the ratings. Hence usability of reputation systems should be given a lot of attention.
April 15th, 2007 at 10:58 pm |WordPress database error: [Table 'i203.is203_users' doesn't exist]
SELECT * FROM is203_users WHERE ID = '1' LIMIT 1WordPress database error: [Table 'i203.is203_users' doesn't exist]
SELECT * FROM is203_users WHERE ID = '1' LIMIT 18. Sean_Carey | April 26th, 2007 at 4:49 pm
I really enjoyed “Writing in the Body: Gender (Re)Production in Cyber Interactions” I find that the internet presents an excellent place to export one’s gender, because of its anonymity. Online, users do not have to be worried about maintaining reputation on social websites such as myspace. I also think its good evidence that gender comes from within the mind, and not the physical features of the body. With the web, you create an avatar to represent who you are in your mind, not body (Most people make it reflect their body because it reflects how they see themselves anyways). I was surprised that O’brien did not go into more detail about this subject. Perhaps because the piece was written before the take off of social websites such as Myspace.
However, on a sad note, we still project our phobias about gender to the online world too. It’s sad that people become offended when they find out that the person they are engaging with are not physically the same gender as whom they portrayed online.
April 26th, 2007 at 4:49 pm |WordPress database error: [Table 'i203.is203_users' doesn't exist]
SELECT * FROM is203_users WHERE ID = '1' LIMIT 1WordPress database error: [Table 'i203.is203_users' doesn't exist]
SELECT * FROM is203_users WHERE ID = '1' LIMIT 1Leave a Comment
You must be logged in to post a comment.
Trackback this post | Subscribe to the comments via RSS Feed