I spend my life answering other people's questions: I'm an Internet librarian. When I discovered the Net back in 1995, I thought it was a great leap forward in the eternal quest for answers. Sure, as long as I used every bit of my information expertise to sort out the helpful from the junk. The Net may be proof of the theory that 90 percent of everything is garbage. But the other 10 percent can save you grief, time, and even your life. As possibly the only living humans who actually read the help screens, librarians have learned how to use search tools really well and have mastered a variety of techniques to root out answers. All it requires is determination—by God, I will find it—and an absolute conviction that every secret is out there somewhere, waiting to be found. In that spirit, frustrated searchers, I am here to help.

Search small pieces of the Web, not the entire Web. Here's a tough question: "I heard some songs by a guy called E. Who's E? What else has he done?" Do you have any idea what happens when you type just E into a search engine? It's not a pretty sight. Deal with this by searching through a subset of the Web, such as a list of rock music performers that can be browsed in alphabetical order. By typing bands into my favorite search engine, the always amazing Google [google.com], I found Artist Direct: Ultimate Band List [ubl.artistdirect.com], clicked on E, and found a brief biography (his real name is Mark Everett), career information (he's now with the Eels), a discography, and links for tour dates, downloads, and reviews.

Use precompiled directories as well as search engines. Why reinvent the wheel? Subject specialists, librarians, and other information professionals have compiled their own directories of links, either exhaustive or "best of," for just about any topic you can imagine. There are even directories of good directories. A fantastic jumping-off point for any search is Lii.org: Librarians' Index to the Internet [lii.org]. Let's say you're trying to find the value of a used car. By checking the Librarians' Index to see which car sites it recommends, you will not only find such standard blue books as Edmunds.com, but will also be reminded to check for fuel economy [www.fueleconomy. gov] and use an auto loan payment calculator [bankamerica.com/ tools/auto_payment.html] while you're at it.

Narrow down sweeping topics. How can search engines help you pare down a broad subject like abortion? They have different strategies. Northern Light [northernlight.com], Vivísimo [vivisimo.com], and the Zoom In button on Excite [excite.com] all organize results into subtopics such as abortion debate, pro-life, and reproductive rights so you can focus your hunt much faster. (We librarians love this kind of thing.) If you're from the "I'll know it when I see it" school, use Google, scan the results for one that's exactly what you had in mind, and click on Similar Pages. Google will reshuffle the deck and generate a set of matching sites, giving you the benefit of both subject-searching and damn fool luck.

Understand that different search words produce different answers. A student once asked me for research on whether obese people earned less than average-size people. There's more than one way to describe each of the concepts in question, so to track down the answer I tried various combinations of fat or obesity or weight, combining them with wages or salary or employment discrimination. On Google, obesity and salary yielded almost 5,000 hits. Strangely, a scan of the first page showed nothing relevant to the inquiry. A simple change to obesity and wages, however, included two spot-on articles—despite a preponderance of health items about "the wages of obesity."

Know when to use your library's Web-based databases rather than a Web search. When somebody says he needs a "study" or the latest "research" on a topic, those magic words send me not to search engines but to the licensed databases of full-text articles that virtually every library offers by way of the Web. The first rule of information is "Go where it is." There is some research on the Web, but most often, research is reported first in journal articles. So for the obesity question, I clicked on my library's Web site, went to EBSCOhost, and searched through journal articles. Check your own library's site to see what full-text databases it offers.

Know what makes information trustworthy. The Net is full of inaccurate, misleading information—or downright lies—so you should always ask, "Sez who?" What are the credentials of the source? Just as important, who sponsored the research? Watch out for hidden agendas, such as might be found in interest groups and corporations. Look for Web sites sponsored by agencies that have reputations to protect: libraries, universities, government agencies, hospitals, and well-known organizations like the American Cancer Society. They're all very careful about getting it right.

Use "wedge words." I coined the term wedge word: a powerful expression you can tack on to your search to help pry out a specific fact or type of document. An excellent choice is FAQ, which stands for "frequently asked questions." People ask the same things over and over (believe me, I know), and experts have posted answers to these frequent queries on almost every conceivable topic. Other wedge words you can add include: research, as in "identity theft research," to pry out original information that meets academic or scientific standards; comparison chart, as in "laser printers comparison chart" (helpful for making decisions, like which printer has the most features you want for the price you can afford); news, as in "Taxol news," important because even reliable sources or pharmaceutical databases may not have integrated brand-new information, such as FDA approval; encyclopedia, as in "art encyclopedia," which virtually guarantees you reliable background information; database, as in "bioethics database," when you want to search inside a collection of data.

Move from general to specific (and sometimes back again). You can do this by using both specific terms and more general ones (salary and wages are specific terms, while employment discrimination is more general). You can also do it by using and—if you're getting too many results, use and's to narrow the search. Think of how you'd choose the perfect mate. You'd start with gender, which reduces possible results by half, and then start anding in other qualities. And American. And single. And between 30 and 45. And Unitarian. And rich. You're now down to about five people. And crazy about you. Oops, you just anded yourself out of a mate, so you go back up the continuum and eliminate one of the and's ("Well, I can be flexible about Unitarian").

Explore the invisible Web. There are answers on the Net that general search engines simply can't find—some sources are fee-based or require registration (you might have to input your library-card bar code to use your library's subscription databases), or you have to manipulate raw data to get the answers. How do you discover what librarians call the Invisible Web? Use guides like InvisibleWeb.com or librarian Gary Price's excellent interface Direct Search [gwis2.circ.gwu.edu/~gprice/direct.htm]. You'll be amazed at what's tucked away. And don't miss hints from Y-Life contributor Danny Sullivan at Search Engine Watch [searchenginewatch.com].

Ask yourself, "Who would produce this kind of information?" Someone once asked me how often teenage boys take showers. (Librarians hear it all.) I knew that the people most likely to compile lifestyle research are those who make money from it: marketers. Since marketing magazines report on surveys of consumer behavior, I went to my library home page; clicked on its business database, ABI/Inform (available at many libraries); typed teenage boys and showers; and found an article in American Demographics called "Squeaky Clean Teens," which analyzed boys' preferences in bathing, deodorants, and hair spray. Would you believe that more than a third of American teenage boys shower more than once a day? Sometimes the Net finds truly surprising results.

Marylaine Block [marylaine.com] is an Iowa-based writer, Internet trainer, and self-described "librarian without walls."