False Feedback Systems and Credence in Online Reputation

I'm not sure I know what the sentence means, but if I understand it overall, you are saying that people too easily dismiss feedback systems because the errors they make appear obvious, but no system of reputation is perfect. If that is indeed what you mean, I don't know what "transparent data doesn't exist" means.

However, because most feedback providers are sincere and because various algorithms can help the owners or users of these sites weigh the feedback provided by reviewers who have proven their reliability, the accuracies in reflections are quite high.

This doesn't make sense to me. Whether people are sincere in the opinions they express in feedback systems doesn't tell me at all whether the information they are purporting to convey is accurate. That there are "various algorithms" applied to the feedback provided is not any kind of encouragement to believe in its accuracy, either. One would be inclined to ask, what algorithms, and why do they make the data more accurate?

Therefore, there exists an inherent whistle-blowing mechanism that will deter excessive investments in reputation – namely the fact that overreliance on particular proxies might contribute to their deterioration as a signal of quality.

Why is this rather obscure proposition a consequence of the prior propositions, which were themselves not too securely put together?

The reader has not yet been presented with any concrete information. The reader doesn't know what the thesis of the essay is overall. But he or she has discovered that the logic presented appears rather shaky and the work of getting the meaning out of the sentences provided is hard. Any such reader, not coerced to continue, will simply give up.

You have to provide us, first, a succinct statement of some idea that will make us want to continue reading. Second, a development of that idea from its essence to some of its consequences and the answering of some likely useful objections, that we can follow without unduly taxing our patience or our willingness to work things out for ourselves. And third, sufficient concreteness, or relation to the world of digital activity around us, that we can judge how you are interpreting the technology, in relation to whatever law and politics you are discussing.

Despite any perceived inadequacies in the above system, the government has declined to become involved in removing or regulating the falsities in these reputational feedback systems. There are market actors with substantial comparative advantages over the government, and they are already doing a fine job of making new information available to the public. In these settings, the government’s role may be most properly confined to facilitating the adoption of uniform standards so that information can be aggregated easily from among a number of different websites, and reputations can be transported from one site to another See generally Nolan Miller et al., Eliciting Informative Feedback: The Peer-Prediction Model, 51 Mgmt. Sci. 1359 (2005)? .

Recent case law has similarly discouraged governmental involvement in the generation of scores or ratings. For example, Browne v. Avvo, Inc. Browne v. Avvo, Inc., 525 F.Supp.2d 1249 (2007)? involved a website that rates lawyers, Avvo.com. The website aspired to do for attorneys what Yelp did for restaurants – provide consumers with information they could use to find a suitable lawyer and collect evaluations from fellow attorneys and clients. However, within ten days of its launch, Avvo was sued in a class action by attorneys alleging that Avvo had violated Washington State’s Consumer Protection Act Washington Consumer Protection Act, Rcw 19.86.020? by disseminating unfair and deceptive information about lawyers who were rated by the site.

More precisely, the complaint faulted Avvo for being subjective and unreliable; providing questionably low numerical ratings to highly regarded lawyers; using a non-transparent methodology for developing lawyer ratings; and providing incomplete information. Plaintiff Browne was an attorney who claimed to have lost two clients due to a poor Avvo.com rating that was tied to his admonition in a state bar disciplinary proceeding against him John Cook, Respected Lawyer Wants Rating Site Avvo Closed, Seattle Post Intelligencer, June 12, 2007? .

Avvo moved for dismissal, arguing that its’ services were protected by the First Amendment and that the plaintiffs had failed to state a claim under Washington’s Consumer Protection Act. The district court granted the motion for dismissal on both grounds. If Avvo were liable for its conduct, then under similar logic, Yelp could be liable for unfavorable restaurant reviews that resulted in a poor rating, AngiesList? could be liable to independent contractors that were poorly reviewed for construction projects and lost future bids for household repairs, and eBay may be liable to vendors that could not make sales due to poor feedback ratings.

Of course, removing the possibility of liability in cases where inaccurate feedback is reported on a ratings website creates the potential danger of diminishing the quality of the published feedback, similar to the risk of declining quality in newspaper reporting that might occur if defamation was eliminated. However, the decline in the quality of ratings and information would not be inevitable; it would depend on whether market forces were in place to provide adequate incentives to keep the information on reputation ratings sites generally accurate.

The court likely made the correct determination in Avvo because individuals like Browne who believe that false information was disseminated about them have a right of reply – an ability to explain why they believe they received inappropriate ratings from a website or a complaining consumer. This right of reply is already built into most consumer feedback systems. Just as Congress enacted §230 of the Communications Decency Act Communications Decency Act, 47 U.S.C. §230? to avoid chilling internet discussion, courts are providing immunity against tort suits stemming from unflattering ratings, so long as the defendant offers the poorly rated individual or firm a right of reply similar to eBay’s. This rule permits a vendor to point out possible biases that formed the basis for an unfair rating.

But this ignores the significance of the First Amendment ruling, which is absolutely straightforward and completely fatal to the tenor of this argument. The holding is, of course, that the website is nothing more than a collection of opinion, protected under the First Amendment precisely because it is nothing but opinion. Judge Lasnik is able to see that the pretense of "reputation" presented by such a site is utter nonsense. Rumor when accumulated is not the same thing as reputation, and Judge Lasnik assumes, perhaps incorrectly in view of the remainder of your analysis, that everyone is able to tell the difference on sober second thought.

The cardinal error in these analyses of feedback "systems" as "reputation markets" is that no one in these so-called markets has to bet with real money. The whole point of the thing is that it's a pretend market in which all transactions are conducted in an infinitely inflatable currency. If, on the other hand, people had to pay money to create these "opinions" or "feedbacks," or whatever, only two classes of commentators would predominate: those who had a significant material incentive to boost the rated entity, or those who had a significant monetary or personal motive to harm it. So the votes cast in an actual "market" would be biased, while those cast in a phony market are simply meaningless.

Marketmakers like these things, because unreliable gossip makes people feel more comfortable in a market, regardless of whether any actual information is being conveyed by the incessant chatter: no one likes to eat in a silent restaurant, and muzak is played in retail outlets the world over. Feedback bullshit is the retailing muzak of on-line selling. "Business intelligence" systems like this stuff, because you can spy on the people who use the forums, thus better "knowing your customer" in an uncomfortably near-Biblical sense.

But, as the shaky logic of your initial exposition tended to show, it's difficult to make a convincing case for taking any of this seriously, ever. If the point of your essay is to make that argument, we need to take a hard look at how you mean to do it. If you wanted the heroic assumption of this dubious premise, you should have made the request at the outset, and showed us clearly what valuable intellectual path would be traveled, to what interesting and worthwhile destination, as a result of indulging the apparent counter-factual.

And in the end, is the real purpose of all this to tell us why government should not interfere in the free exchange of gossip and unsubstantiated opinion? Against the background of the First Amendment, this is—as the judge says quite vigorously in the very case you seem to rely upon—this is all but tautological.