Friday, March 18, 2005

"Getting Thousands of Hits" - Ban This Meme

Our colleague Lorcan Dempsey's post to his blog yesterday was about library strategic plans and searching from library web sites. One commenter to Lorcan's posting worries about (my words, not his) dumbing down the library's search interface to the simplicity offered by Google, fearing the result would be "thousands of hits."

I have often heard this sentiment expressed by librarians with regard to ordinary people using generic search engines. Indeed, the president-elect of ALA did that in his piece on Google Print. (abstract only here...spend $$ to see the whole article) The notion is that, because of the size of the web, using simple search terms, or imprecise ones, yields way too much information, undifferentiated and untethered from helpful structures such as LCSH.

I am announcing my official renunciation of this needlessly binary view of searching. It's unhelpful, breeds complacency, and is myopic.

Pretend you've never ever been in a large library. Pretend you know absolutely nothing about taxonomies. Pretend you don't know the difference between a magazine, a journal, an index and a book. Pretend you don't know what you don't know, and don't know how to articulate your unknowingness. Once you've pretended all this, make a pretend visit to a very large library for the first time.

And then tell me why it would be easier to find information in this situation than it would be to find information in a set of Google or Yahoo! search results? If you're pretending to be in a physical large library, maybe you'll ask for help, but there's a good chance you wouldn't. And if you're pretending to visit this large library virtually, chances are you can't see any human presence at all. In both places, the pretend you might wander around a bit, poking this button or looking at that printed object. But in the absence of any knowledge of how the "thousands of hits" in a library may be found, let alone investigated, the ordinary searcher might leave the library with nothing. A scarcity of information.

Yes, people get thousands of hits when they search Google, and they do because it's easy. But, it is a convenient fiction librarians tell themselves that nothing of any use is found this way because of the abundance of information. It just isn't true. I say so in the introduction to The Environmental Scan: almost all of the 250+ references in the bibliography were found using Google, using keywords. So, let's get past this unhelpful "us vs them" mentality and work a lot harder at making the valuable information we curate easy to find.

In November 2004 I blogged about Google Scholar and wrote about a post from a blog called Dog News. I said: "This is how scholarly material is going to come to "the people" rather than the people coming to scholarly material which is the current paradigm at work in libraryland, and one that consistently fails to be attractive to searchers."

Dog News caught my mention and commented on it, saying, in part, this:

"I GET this, why not make a real world example of it, so everyone else, those of us who aren't brilliant librarians, can GET it too. Google Scholar can change the world! and I can help to point the way... knowledge is power, leveraging knowledge to your goals is even more powerful."



stevenb said...

To argue that Google is a useful way to find information is situational. It depends what you're searching. If you want quick reference, facts, the latest dirt on celebrities, background information on an artist, the reasons for the civil war - yes, Google can provide useful information. At other times, for more complex research needs (I helped a student the other day who needed detailed information on how developed nations are preparing for new epidemics - and he needed some types of information that Google could not satisfy) Google's thousands of hits are useless and wasteful of time. I can agree with your "us vs. them" concerns. I'm call myself a "resistor". I not only believe that most library resources are already simple enough (most users I know figure out basic search screens easily), but I oppose suggestions that all library resources should be like Google - or similarly converted to all purpose metasearch interfaces. I think we have a lot to lose if we go in that direction. What I do support is user education. There is a new myth that anyone who supports user education is out to turn everyone else into a librarian. That's absurd. What we are trying to do is enlighten the user community to their options. Becase all search is situational, you need to know when the correct source is Google - and when you need, for more complex searches - a higher level aggregated database. And the simplisic "google vs. library database" ignores other complicating factors. Sometimes mega-fulltext library databases can give as much poor information as google can - if used poorly and with poor search strategies. For many disciplines you absolutely need to have the common sense to search a specialty database. An architect must know when to consult the Avery Index, just as a finance major must know when to consult compustat. Where will students learn this? From librarians - and hopefully - not too far in the future - from their instructors. I have read the section of the Environment Scan that relates to individuals preferring the simplicity of google. This fit well with those who are members of the "good enough is OK" school of thought that suggests that we should expect today's information searcher to try to do better than getting a couple of "good enough" sources from Google - and even Google Scholar does not necessarily improve on "good enough". I can do a search on information literacy and easily demonstrate that while I will find some relevant resources - there are many recent, excellent articles on information literacy that are missing - and that's what I think it is about - not "getting thousands of hits" but "it's not about what you find - but about what you miss." That's what I try to educate users about.
I hope that we can all, those who prefer "googlelizing" and those who "resist" can pay more attention to the value of user education - especially when it is embedded into all curriculums. For more information read my "Infodiet" article from the 2/20/04 Chronicle of Higher Education - and the summary of the "Googlelizer vs. Resistor" debate from a December 2004 issue of Library Journal.

Fichter said...

Are we dumbing down something or smarting up the interface? I happen to think Google is a smart interface and there is a lot of paddling going on furiously under the hood that I see in order get a clean easy to use result list. Depending on research question, Google can be the best choice or the worst. We can stack the decks either way.

I agree with Alane that when someone approaches a large academic library and they don't know where to start or what to search. In fact, that often the task is daunting and fraught with failure when they do attempt it. Look at your OPAC, site search and database logs to see the evidence of this confusion. When they try to choose a database in a subject area unfamiliar to them, they don't take time to read the database descriptions - they pick the first likely looking one on a page of databases, often the very first one in an alphabetic list of databases, or they choose their tried and true database that worked before for a entirely different discipline.

We've done several rounds of usabilty testing in the past year looking at how people choose databases and how to ameliorate the high failure rates of users trying to find articles on academic library web sites.

Federated search would be one way to help reduce this failure rate and would be a great place to start (a smart choice) for many of our users faced with choosing from a array of databases with cryptic brand names. Ideally the federated search results can have an interim page that reveals and expose what databases are being searched. That's a "teachable" moment for those that care to learn more.

Should a federated search be the only choice? No. But I don't see any serious supporters of federated search, saying we should toss out course pages, subject guides, tutorials, and other resources that inform users about resources.

I would echo what Steven that teaching is important, in fact, it's mandatory given the high failure rate of students to find articles on academic library sites ("Serial Failure” The Charleston ADVISOR, Vol. 5., no. 3, 2004. Jennifer Bowen, Judi Briden, Vicki Burns, David Lindahl, Brenda Reeb, Melinda Stowe, Stanley Wilderbut). The reality is most users come to our site for "self discovery" and are in "self service". We have to be prepared serve them accordingly. Our sites need to support independent self discovery while offering the option to learn more.

Peter Pirolli and Stuart Card's work on information foraging theory gives us insights into how people hunt for information and make decision whether to continue or stop. They noted that our patterns are similar to animals in the jungle hunting for food. Jakob Neilsen commenting on their work noted that users like to catch small rabbits. Why? They're easier to catch and it looks like it will take less effort. Users satisfice all the time on their hunt for information and we have to design smarter systems and better web sites that surface the information that they need quickly. It's time that library web sites made our content look "easy to catch".

Anonymous said...

fichter is damn right!!!

Anonymous said...

"To argue that Google is a useful way to find information is situational. It depends what you're searching."

Like News & Blogentries, which can be found better by using a blog search engine like plazoo or feedster

Anonymous said...

How does search engines like plazoo or feedster work?

Anonymous said...

How does search engines like plazoo or feedster work?

Anonymous said...

Nice question. I daresay someone's out there to give an answer?

Anonymous said...

There you might find your answer !