There isn’t a better time to discuss information-processing biases than during the Discovery Channel’s Shark Week, which started yesterday. A common cognitive bias is overestimating the risk of shark attacks. Because they are heavily reported, in a manner that tends to elicit strong emotions, we think they are more common than they actually are. This misconception is sometimes called “availability bias”, or the “availability heuristic.”
I released a paper today, titled “The Search Fixation: Infringement, Search Results, and Online Content,” which relates to another type of information-processing bias: the “streetlight effect.” Sometimes called the “lamp post effect”, this bias is usually attributed to researchers, and refers to biases resulting from data-gathering where the data is most easily gathered. The principle is best illustrated with a joke related by David Freedman in Discover Magazine:
Late at night, a police officer finds a drunk man crawling around on his hands and knees under a streetlight. The drunk man tells the officer he’s looking for his wallet. When the officer asks if he’s sure this is where he dropped the wallet, the man replies that he thinks he more likely dropped it across the street. Then why are you looking over here? the befuddled officer asks. Because the light’s better here, explains the drunk man.
We look where it is easiest to look. By definition, one of the easiest places to look for something online is a search engine. If one uses terms like “torrent” and “free download”, it is possible to find search results that linking to infringing content. Because search results are easily observed, the streetlight effect tells us, this is where the observation occurs. As a result, some will fixate – like this RIAA memo – on infringing search results, notwithstanding the fact that pirates tend not to use general-purpose search engines for finding infringing content.*
Consider an example used in the RIAA memo: “free bruno mars mp3.” While this result may return results that point to infringing content, data from Google Trends indicates that strikingly few people are actually running that search. A worldwide tally of search results through last week, comparing “bruno mars” (for reference) to “free bruno mars mp3” is available here. Bruno Mars is orders of magnitude more common, and at periods of time, no one seems to have searched for “free bruno mars mp3” anywhere on the planet. This is not entirely surprising: extensive research by BAE Detica indicates that the majority of infringement “business models” don’t depend on search engines at all. Sites like Isohunt and the Pirate Bay shrugged off   the threat of search demotion, announcing that they receive little traffic from search engines.
Because it is easy to test search results, however, search results are where testing occurs, and that can lead to inefficient decisions about enforcement and policy priorities.
* Note that regardless of whether search engines are important to infringers, search engines need to comply with requests to take down content upon an infringement allegation, in order to maintain certain liability protections: this is required by Section 512 of the Digital Millennium Copyright Act.