Know Your Results Before You Search

Imagine going, money in hand, to a storefront that gives no clue as to it’s merchandise. Will this store sell shoes, food, antiques, or pets? Lined up next this blank storefront are about 50 others, each with a more or less obtuse name, and each with a microphone protruding from it’s door, ready to receive you requests. What would you do?

This is what students face when they first delve into online searching. From long familiarity, most students have a pretty good idea that they’re searching the author’s text of web sites when they choose to request information from Google. Some might even know that authors can supply metadata, but many do not. And most are familiar with the result list as it is presented, and what they can expect to see when they click into a result.

But what about library databases? Some allow full-text searching, some only include citation, abstract, and controlled vocabulary. As the librarian for languages and literature, my primary tool is the MLA International Bibliography (Chadwyck Healey), and this database doesn’t even include an abstract, and the controlled vocabulary is not very evenly applied. (What I would give for MeSH-like terms!) But my students also need to know how to manipulate catalogs and several other databases, including JSTOR, which replaces MLA’s subjects with fully searchable full text. There is almost nothing similar about searching these two databases because everything inside the storefront is fundamentally different. The words you choose, the care with which you construct searches, the results you get… everything changes.

So I’ve begun teaching databases backwards. I start with a record and get the students to understand what’s there and what’s being searched. Then we move back to the result lists and figure out how to leverage a “messy” keyword search by analyzing the results, quickly opening anything that looks remotely relevant, and gleaning search terms that way. Then and only then do we move back to the search interface and I’ve found it takes very little time for the students to understand what’s going on now that they know what they’re searching and what they can conceivably expect to see when they click “Go.”

My favorite question from a student came part way through one of these backwards searching experiences. “Can I use these same ideas in ProQuest?” I answered that she could use exactly this technique to learn at least the basics of any of our databases. But I knew she still had more to ask… “Can I use these same ideas in Google?” She asked. Yes — yes you can.

I need to plant that kid in all my classes.

[Update: Thanks to Mark for pointing out that as I've laid this scheme out here, it works best in class situations. But with a little tweaking it also works in "real life" when you confront a new database. See the comments section beginning here for that exchange.]

10 thoughts on “Know Your Results Before You Search

  1. Thank you Iris. Both the “storefront” idea and the “teaching backwards” idea are spot on. I’m eagerly anticipating the next “how do I search databases” question I get at the reference desk…I can’t wait to try out your backward technique.

    I’ve just circulated your post to our reference librarians. Thanks again.

  2. How nice of you to say, Kathryn. I hope the idea sparks more ideas among your librarians. And please let me know if you and they come up with ways to reinvent it or make it more effective.

  3. Hi Iris,

    This sounds like a wonderful idea to me, and bless that young woman! But I have a procedural question. If one is using a new database for the first time and wants to use this technique, how do they get a record in the first place?

    Do you recommend that they run any sort of search to get any record, or is there some other technique that I’m missing?

    I do think this is wonderful since although I may have never articulated it, I have often wondered how I am to (effectively) search for something when I have no idea what form it might take.

  4. I usually start with a keyword search for “frogs.” It’s very scientific… What’s even more amusing is that I’ve never run across a databases that doesn’t return at least a couple of hits mentioning frogs.

    But in a class situation I usually do this using screen shots so that I can move everything along without muddying the waters with “now let me just get you to a record in JSTOR for comparison…” Then we move to live search as soon as they’ve explored the screenshot of a full record. So in these situations I usually make screenshots based on searches relevant to the class’ paper topics (if I can divine them).

    At the reference desk I mix the first two steps together. So the student and I do a “messy” keyword search. Then we take some time to analyze a couple of records and the result list, gathering terms along the way. Then we experiment with various advanced techniques based on the terms and the record structure we’ve found.

  5. This sounds like a good strategy and I assumed you had a technique for class and at the reference desk, but I fear I wasn’t clear enough.

    And, yes, frogs are scientific and cool!

    What I wasn’t too clear about is what strategy do you recommend to get that first record when someone–without the expert librarian by their side–encounters a new database for the first time? I assume you are are also trying to teach appropriate self-sufficiency for lifelong info literacy.

  6. When I encounter a new database I’ll do a super-dumb keyword search for “frogs” (or anything else that strikes my fancy), I open the first one or two records and look carefully at them. Is there an abstract? Is there controlled vocabulary? Is there full text? If there is full text, does it look like my search was run on the full text as well as on any citation, abstract, or controlled vocabulary that might be present? (If I can’t tell about the full text searching I’ll take a long-ish phrase from the text and run a phrase search on it to see if that record reappears.) Are there lists of works cited and if so, did my search get run on these lists as well as everything else? Is there some field unique to these one or two records that I might be able to manipulate?

    So I guess what I’m saying is that by having a dummy search that I can routinely do on new databases I’ve freed myself from the compulsion to actually manipulate the database and ask it for what I really need. I’m simply trying to get something, anything, that’ll help me figure out what types of searches I can do when I return to the search screen to hunt for my actual information need. I’ve gone up to the storefront and asked for “stuff” so that I can figure out if this one sells food or pets.

    Does that make sense? In other words, when I’m using this technique in “real life” rather than in classes I can’t ACTUALLY skip the first, frightening search screen. All I can do is sap it of it’s scariness until I can figure out what it can actually do for me, and how I can have a hope of getting it to do what I need.

  7. Yes’m, it makes perfect sense. I was only trying to get you to lay it out in full for those who might need it made explicit. You know, people like me. ;)

  8. Hi,

    I just wanted to comment and say I love this idea of teaching ‘backwards’, which once you explain it makes a lot of sense! I’m hopefully going to give it a go in my conservatoire’s class resource training sessions this autumn – as a music librarian I think my equivalent of ‘frogs’ will be ‘Mozart’ or ‘concerto’ or something, but the principle still stands :)


Comments are closed.