A Survey of Internet Searches and Their Results
Andy Barnett · Reference & User Services Quarterly · 1999
In this study, we collected more than two hundred Internet searches using Magellan Voyeur, and then reentered searches in Magellan and Hotbot. We categorized searches by search operators used, including search engine math, phrase searches, Boolean logic, natural word searches, and Web address searches. Here we assess relative success of various strategies, as well as implications for user education. The Internet has become pervasive feature of modern life, with more than ninety million users in North America alone. It is part of curriculum in most elementary schools, widely available free at public libraries, and major feature of pop culture. Commonly touted as an electronic library, it is also derided as stuff swamp, the world's largest pamphlet file, or a library assembled piecemeal by pack rats and vandalized nightly.(1) When Internet is considered an entertainment medium, difficulty of finding specific item on Internet might actually be an attraction to users, Web owners, and advertisers. Users find occasional gold nuggets of entertainment and information while they're searching for something else, and psychologists, gamblers, and lab rats will tell you that random positive reinforcement is most powerful kind of incentive. Site owners know that being entertaining or merely bizarre can generate traffic and consequent advertising revenues. Advertisers seek user's trancelike state that starts with few clicks and ends hours later without sense of time's passage, whether this occurs in front of computer monitor or television. When Internet is considered an information medium, extent that users are able to find what they are seeking is measure of Internet's usefulness. About 85 percent of users go to search tool when they want to locate something. Search tools are some of busiest and most profitable sites on Web. What methods searchers use and how well various methods work has implications for those instructing Internet's many new users. In this survey, we look at actual searches to determine what kinds of searches users are performing and how successful various types of searches are. Methodology We collected total of 207 searches in March 1999 using Magellan Voyeur, which displays random searches currently being performed on Magellan site.(2) The aptly named Magellan Voyeur allows viewer to follow any of displayed searches. The user can also copy and save displayed searches, as we did for this survey. No identification of original searcher can be made. Though searches displayed are random, searches in this sample were not. To minimize searches from classrooms and businesses and maximize searches by individuals, we gathered data on weekdays between 7 and 10 P.M. (CST) and on Saturdays between 5 and 6 P.M. We only included English language searches, as foreign language searches are difficult to evaluate for spelling errors. Since Magellan is not heavily advertised and does not command large market share, users of Magellan site have sought out well regarded but not well-known tool. While we can not identify searchers in study sample, time and language constraints make it likely that they are from either United States or Canada, are not at work or school, and might be slightly more computer savvy than average. In this study, we reentered sample searches in Magellan and Hotbot, and noted total number of hits returned. We categorized searches by search operators used and noted success of various methods. We also analyzed effect of obvious spelling errors. Magellan, like better-known Yahoo!, is subject directory that selects sites and arranges them into subject categories.(3) Yahoo's selected sites and categories are searched, as opposed to search engines such as Alta Vista or Hotbot, which try to cover wider portions of Internet. …