II
Google doesn’t know it all

The Google “answer machine” vs. library reference services
Google’s limitations

The Google “answer machine” vs. library reference services

Google has stated that it is “the closest thing the Web has to an ultimate answer machine” http://www.google.com/press/press.html.

In an article in the 25 August 2003 issue of USA Today http://www.usatoday.com/tech/news/2003-08-25-google_x.htm, Google co-founder Larry Page is reported as saying that he still wants people to frequent libraries, even if it is “a lot easier to go to Google first”. This strikes one as a rather uninformed and condescending statement and gives the impression that Mr Page doesn’t seem to have any real grasp of what today’s libraries can offer in terms of information resources, and that there are hundreds of databases, not to mention print and archival resources, which will not be accessible via Google.

However, Larry Page’s statement is in some contrast with one made by Craig Silverstein, Google’s Director of Technology, during a keynote address at the InfoToday Conference 2003 in New York, and as reported in Library Journal http://www.libraryjournal.com/article/CA299471. The ultimate goal of Google, according to Silverstein, is “to seem as smart as a reference librarian”, but he also acknowledged that the Holy Grail for Google – to act like a reference librarian in responding to search queries – was still “hundreds of years away”. He recognized that librarians and information professionals still have a vital role to play in assisting people with their research and information needs.

In a more recent pronouncement, in an interview with the Google founders in the September 2004 issue of Playboy (see The Google founders interviewed), Larry Page says “Libraries might have some of the information but probably not all – and not necessarily the most up-to-date”; Sergey Brin says, “more and more librarians love Google … Finding information still requires skill. It’s just that you can go much further now. Google is a tool for librarians just as it’s a tool for anyone who wants to use it.”

While, happily, reference librarians have not yet been replaced by Google or other search engines, there is no denying that the trend to provide networked reference assistance has had a significant influence on the traditional role of reference librarians at academic institutions. There have been changes in the use of library services for enquiry and reference assistance. Traffic at reference desks in some libraries is reportedly on the decline, with many library patrons preferring to submit questions via e-mail. An “Ask a Librarian” electronic reference service is now part of most academic libraries’ Web sites, and librarians are increasingly devoting more energy into developing online help.

In a thought-provoking article by Gary Price, “What Google Teaches Us that Has Nothing to Do with Searching”, in the online publication The Searcher (vol. 11, no. 10, November/December 2003), http://www.infotoday.com/searcher/nov03/price.shtml the author suggests that the library community is just as much part of the Google craze as the general public, and that many people have a “Google or bust” mentality; and that while this “Googlefication” was going on, information professionals, as a group, didn’t speak loudly enough to tell people that Google is just the tip of the iceberg, and that Web searching has many limitations, depending on users’ information needs. He goes on to say, “we know that the Web is a tool, not a solution. Have we demonstrated this fact to our users? Not well enough.” While not advocating that Google needs regulation in any way, shape or form, the author says “I do want information professionals to know what choices exist and share them with patrons. Choices and options have always played a major role in what we do.” Price wants reference librarians to become more proactive and able to demonstrate to their patrons what they and their library can do, that there are things that Goggle can’t do well, or at all; that librarians can offer their skills in making choices on where to start searching, on information authority, and on quality issues; and that they can save people time in acquiring accurate and timely information.

There was also an interesting recent study by four Cornell reference librarians, “Google Meets eBay. What Academic Librarians Can Learn from Alternative Information Providers”, in D-Lib Magazine of June 2003 (vol. 9, no. 6), http://www.dlib.org/dlib/june03/kenney/06kenney.html in which the authors seek to demonstrate what librarians can learn from Amazon, eBay and Google, and Cornell University’s free e-mail reference service is compared with Google Answers. In the Cornell study, 24 questions – representing a range of research questions, from simple to complex, typically encountered by reference librarians – were submitted to reference staff at Cornell and to Google Answers. The answers from both services were then evaluated on a blind review basis by university librarians, who were asked to grade the responses as poor, fair, good, very good or excellent.

The study found that librarians fared better overall, but only marginally. Although Cornell reference librarians scored higher overall than the freelance researchers for Google, their scores were not significantly better. The authors recognized the limitations of their study, and that it was more exploratory than scientific, but felt that it nevertheless provided potential lessons about how the commercial sector handles market-driven reference services, and what academic libraries can learn from them: “Academic librarians must become more savvy in articulating their value to the educational enterprise in order to prosper in a rapidly changing information environment.”

These views are echoed by Gretchen Walsh, Head of the African Studies Library at Boston University. In an essay identifying obstacles to successful research on Africa in a forthcoming issue of The Reference Librarian (Haworth Press, volume 42, issue 87/88, November 2004), she says “Reference librarians must do all we can to reverse the tendency for students to bypass much-needed reference assistance, and we must do all we can to make their research tools work for them, with or without our guidance.”

Pitting Google’s Web Search to the test against more conventional methods of finding quick access to facts is now becoming something of a sport. For example, the UK Guardian newspaper, in their issue of 6 May 2004, “On your marks, get set, search …”, accessible in the Guardian’s online archive at http://www.guardian.co.uk/g2/story/0,3604,1210327,00.html, conducted an interesting experiment: it set six questions comparing Google searching against other methods of finding information quickly, by telephone, and via library reference research. The questions were of a general nature: (i) to identify the books written by the editor of a well-known UK tabloid newspaper, (ii) to find passages from speeches or “rants” made by former British Prime Minister Margaret Thatcher, (iii) to track down the name of the Vice-Chairman of the UK all-party parliamentary group on back care, (iv) to find out what proportion of the Slovenian railway is electrified, (v) to discover what Sophie and Edward (two members of the British royal family) did on a certain day, and, finally, (vi) to find out what was unusual about the British gold medal victory in the 400m race in the 1908 Olympics in London. Google came 1st for three of the questions (although with a couple of reservations on accuracy or currency), 2nd for two questions, and 3rd for one question (on the parliamentary group on back pain). Library research also came 1st for three questions, 2nd for two, but was disqualified on one further question on account of outside assistance. The method of phone enquiries to track down answers came out last by far, as the least speedy. Overall for Google vs. library research, the results were pretty even, but Google’s performance is quite impressive, albeit for answers and quick access to facts to questions of a more general nature rather than academic research.

One thing is clear in any event: the Internet cannot replace the academic library as the primary source for research. In addition to print and archival resources, including indexes and bibliographies, libraries can offer a huge range of electronic databases, most fully searchable, which are accessible to students and academic staff from any computer around the clock but not via Google or other search engines.

 ^top

Google’s limitations

The methods that students and academics use to look for information, and to get quickly to sources of interest, have changed quite significantly in recent years. There have been changes in access and delivery methods, although at this time most research still starts, and ought to start, with keyword searches of indexing services, online library catalogues, and electronic databases. Web search engines, and especially Google, can sometimes also generate excellent results, and are now well recognized as significant research tools across all disciplines. Google can do simple or more advanced searches quickly, and can come up with remarkably good results. One downside of this is that some students, impressed by Google’s capabilities and sophisticated search offerings, tend to think that the extent of their research need not go beyond clicking on to Google.

Nowadays most researchers will include the Web as part of a research strategy, but much will depend on the nature of the research. There are occasions when Google can do the job best, but while Google’s search results can be a very helpful gateway and starting point for tracking down information, one must always be aware of its limitations. Google doesn't know it all, and any researcher must always apply careful critical judgment in the choice and evaluation of links to information and Internet resources which Google, or other Web search engines, are able to provide. (See Section V Sample search strategies – Evaluating success and relevancy ratings.)

The first thing to remember is that, at this time at least, Google searches are limited to publicly available network resources on the Web and the “open Web”, also called the “visible Web”. While Google lets you search freely accessible pages on the Web, including Word documents, the text of most freely accessible PDF files, and files in other formats such as PowerPoint or Excel, the Google robots cannot reach, or are prevented from reaching – and are therefore unable to index – password-protected online resources and databases that are available only by subscription or under licence, and/or require registration before access is granted. Additionally, there are Web pages and databases, often containing high-quality and authoritative information, which search engines will not index for technical reasons, or will not add to their indices as a matter of deliberate policy. These are the online information sources that are part of the “invisible Web” or the “deep Web”. The BrightPlanetÔ Corporation has published a listing of the 60 known, largest “deep Web” sites http://www.brightplanet.com/infocenter/largest_deepweb_
sites.asp that contain data of about 750 terabytes, which amounts to roughly 40 times the size of the known surface of the Web. These sites appear in a broad array of domains, from science to law to images and commerce. BrightPlanetÔ estimates the total number of records or documents within this group to be about 85 billion, and it is constantly growing.

This invisible Web includes not only a vast number of proprietary (fee-based) databases and online information services, but also numerous openly accessible government and public records in digital formats, statistical and image databases, archives of newspapers and magazines, digital library projects, as well as, for the most part freely accessible, content-rich databases from libraries and educational institutions around the globe.

 ^top

For example, in the African studies field, Google does not index searchable databases such as David Bullwinkle’s massive Bibliography of Africana Periodical Literature Database http://www.africabib.org, an English-language database that indexes over 50,000 articles and contains citations from over 400 African and development-studies periodicals and other journals that cover Africa on a regular basis. Similarly, it does not cover indexing services such as The Africana Conference Paper Index http://www.library.northwestern.edu/africana/resources/index.html, an important and substantial index, updated daily, to the individual papers of the Africana conference proceedings held by Northwestern University Library in western European languages.

Moreover, as indicated above, any Web site that requires a user name and/or password is out of bounds to search engine robots, and that includes a very large number of scholarly journals. Thus Google’s ability to track down and index articles in African studies periodicals is strictly limited: it will have indexed articles in freely accessible online journals, but not material which appears in the leading, commercially published African studies journals, although it will pick up anything that is freely available on the Web sites of the publishers of these journals – such as general editorial information, tables of contents of current and back issues, and abstracts of journal articles; but access to full-text of the articles themselves is usually subject to subscription, or payment of a fee.

As another example: Google will have indexed the pages of the African Studies Companion: A Guide to Information Sources that can be accessed for free, namely the preface, table of contents, and the Introduction at http://www.africanstudiescompanion.com/intro.shtml, but not the listings and resources contained in the database itself, which require subscription access.

The second important point to remember is that although Google is clever (as are some other sophisticated search engines) none of them will interpret your question. Part of a reference librarian’s training is in interpreting the questions posed by enquirers; as a librarian friend put it to me recently, “I can remember very few enquiries that did not require me to ask further questions before I knew what was really wanted and therefore which sources to consult.”

Google says it believes in instant gratification: “You want answers and you want them right now” it says in its Google Ten Things credo http://google.com/corporate/tenthings.html. But interpreting Google search results as “answers” could be fallacious and merely amount to a quick-fix solution. The fact is that Google will simply find the words or terms you are looking for on Web pages. The links it finds and the search results it generates are not the same as answers, much less authoritative and reliable answers, although it might eventually lead to answers or authoritative sources. But Google does not, and cannot be expected to, either interpret search or enhance the information it tracks down.

For some basic questions – for example, brief factual information and statistical data, current population figures, GDP, literacy rates, life expectancy, etc. – consulting an authoritative and current print or online resource may well be quicker than conducting a search in Google. It is therefore important to have at least broad knowledge of the major reference tools and information sources that are available in your university or college library, in both print and electronic formats. This includes the major current bibliographies and continuing sources, and the guides to sources (e.g. bibliographies of bibliographies, directories, encyclopaedias, handbooks, etc.), thus enabling you to quickly identify the most appropriate research tools for finding answers to your questions. There are annotated listings of some of these sources in Section 2 (“The major general reference tools”) of the African Studies Companion, although they are restricted to general and multidisciplinary sources.

This could also apply, for example, when searching for basic information about African governments and government agencies – including current heads of state, ministers for the different government ministries, specialized agencies, etc – which can be quickly tracked down by using some of the online resources listed in Section 1 of the African Studies Companion (“General online resources on Africa and African studies, and the best starting points on the Web”).Moreover, for more general lines of enquiry relating to African studies – for example, the major publishers, vendors, journals, libraries, associations, organizations, online forums, degree programmes, etc. – it will probably be quicker to consult either the print or fully searchable online version of the African Studies Companion, or some of the best African studies “mega sites” and gateways such as Africa South of the Sahara – Selected Internet Resources(entry 47 in the African Studies Companion) http://www-sul.stanford.edu./depts/ssrg/africa/guide.html or Columbia University Libraries – African Studies Internet Resources (entry 56) http://www.columbia.edu/cu/lweb/indiv/africa/cuvl.

And, finally, always remember there is life beyond Google: for any more complex research questions consult your reference or specialist subject librarians, who have the knowledge and skills to put you in the right direction, and they can facilitate access to authoritative and trusted sources both in print and digital formats.

 ^top

Using Google for African Studies Research is  published as an adjunct  to The African Studies Companion: A Guide to Information Sources New 4th edition 2006 now published

 

 <Previous | Next>