Finding information on a topic

Previously, I had a post about finding information in books using things like Google Book Search. This post talks about finding information on a topic, or more specifically, why you should start your search with a research database and more about what research databases are (like the real ones). In a post coming up, I'll give some information on some free to you research databases (the real ones).

You should start your search with a research database to be more comprehensive, to cover multiple sources and publishers, to have real searching power/precision, and because of the vocabulary problem.

We know what research databases cover - they tell us the list of conferences and journals and the years covered, unlike the mystery meat that is Google Scholar. The biggest of the databases, and the most incredibly expensive (well into the 6 figures for an R1 in the US for a single database with a limited number of concurrent users), are pretty much comprehensive in coverage. Take Chemical Abstracts, for example. Pretty much anything you'd want in chemistry. Same with PubMed which is free to you. Even if you're talking about a much smaller database like Aerospace & High Technology, it covers multiple publishers' stuff, government stuff, and stuff in a bunch of different languages. Oh, and the different languages - well they're still indexed in English. If you decide you need it, well then you'll need to find a translation or just deal with the pictures and equations.

If you go directly to a publisher's digital library, IEEE Xplore or Science Direct, for example, instead of Inspec or Compendex, you'll miss things from the other and from Springer, or SPIE or OSA or ACM or any other publisher. And you won't have much power in searching/

Let's talk about power. Analytical abstracts lets you say if the chemical you're looking for is an analyte or matrix (very handy!). Inspec lets you look up frequency ranges (numerical indexing). BIOSIS lets you look up taxonomic terms (like kingdom, phylum, class, ...). A bunch of them let you look for a treatment - application, theoretical, etc. You can be very precise.

The vocabulary problem is basically that different people use different terms to talk about the same thing. Sure, if you work in the area you'll know the difference in connotation between lidar, ladar, and optical radar, for example. But really, articles about any of these might be useful. If you use Google or do natural language searching in another digital library, you'll need to OR all of these variations. Can you think of all of the variations? What about British spelling or American spelling?Real databases pick a preferred term - not one that is better, but pick one to stand for the concept. So you can find out what this term is and pull up all the articles on that concept, even if your word doesn't appear in the title or abstract.

When I say pick one, I mean specialists spend time coming up with a controlled vocabulary and decide what terms are preferred (because they're most common or whatever). These are also arranged hierarchically so you can explode a search which means you can search for anything under that higher level topic. You can also find related terms.... Oh, when I say pick one, I also mean that most of the research databases have human indexers. It's probably machine-aided (machine suggests terms), but there are humans writing the rules and doing quality control. Humans who ask themselves: what questions might this article answer? what queries does this answer? how can I best describe this content so that it can be found by people who need it?

Ok, now you're sold. How do you find one of these bad boys? If you're in biomed (which it seems like everyone online is sometimes), then just use PubMed and supplement it if you're at a research institution with Embase, BIOSIS, Chem Abstracts, or if you're in biomedical engineering with Inspec and Compendex.  For the rest of the world, check out the recommended databases in the subject area on a library's research guides. Pick a library! Or, you can look at the descriptions on DIALOG or STN - they have transactional pricing for access to databases.

I'll tell you about some free ones in an upcoming post.

PS - I almost forgot my example: flutter. Look it up in an aerospace database if you want it for missiles, in a biomed database if you want hearts, and who knows what you'll get in Google!