How can social scientists learn database skills?

December 24, 2005 9:40 AM

What resources exist for non-techy social scientists who want to use relational databases to store their data, but don't know how?

I am a computer guy who has been working extensively with a couple of psychologists on a research project. I knew they were more comfortable with single-table (flat-file or spreadsheet) data representation for data storage, but in this case their study generated much data, and so highly-related data, that I convinced them to give a database a shot. Long story short: half a year later, they're still a little mystified by SQL and wouldn't be able to design their own table schemas if their lives depended on it, but they see the very strong advantages databases have over spreadsheets for their needs. Furthermore, we're writing a paper about how to replicate the study we did, and we want to include a bit about why we used a database instead of a spreadsheet and why we think anyone replicating the study should consider doing the same thing.

One important part of that would be pointers to books and articles that interested readers could look at to learn how to set up and use databases for themselves, which brings me to my question. These interested readers would all be psychologists, and therefore probably not very technical people who'd be comfortable with reading "Theory of Database Design" or "Teach Yourself SQL in 24 Hours" and extrapolating for themselves how that applies to them; they'd want something that explained things in a way much more directly relevant to their work. The problem is, in digging around the Internet, I have found basically no information about databases tailored to people doing social science research. This surprises me; it seems to me that social scientists' problems are often best solved by databases, but apparently they never use them. So, question: are there any good resources (especially books or articles) about how to use a database that are written for social scientists? I'm interested in anything database-related: how to set up a database effectively; how to use SQL; how to pick the right RDBMS for you; anything, really, so long as it's about databases and aimed at social scientists.

posted by jacobm to computers & internet (7 comments total)

What are they doing that's so social sciences-specific that they can't learn basic database concepts from the same places that we learned them? It's not really an issue of technical competence (unless you're trying to throw SQL syntax at them from day one, which I'll admit is a bit scary) -- learning how to design and use a database effectively is really just a conceptual leap.

It sounds like they're more put off by all this talk of SQL and 'schemas' and other technical language, rather than by the prospect of putting some effort into learning a new and useful skill. They were obviously capable of figuring out how to use a spreadsheet to represent their data abstractly in rows, and I bet they didn't need an Excel tutorial that was tailored for social scientists. Couldn't you just set them up in Access or something that has a nice visual interface, and use a standard database tutorial that has plenty of real-world examples?
posted by chrismear at 10:01 AM on December 24, 2005

Microsoft Access has some easy to use GUIs to design relations and whatnot. It was designed for "non-techy" people. That's probably the easiest RDBM to use.
posted by Paris Hilton at 10:05 AM on December 24, 2005

I don't know of any resources offhand, but I'll bet explaining (and showing them examples of) entity-relationship diagrams will help immensely, especially if they're visually inclined. It's a lot easier to figure out what a database does when you walk through real-world examples on paper.

Though not immediately applicable to social sciences, it occurs to me that DB tutorials for web designers might fulfill some of your needs; since practically anyone can make a website, such tutorials are often written for a layperson audience. This one, for example, is a fairly brief but comprehensive introduction to what a database is and how it works from a conceptual point of view. Tutorials aimed at novice web developers are also likely to be far more plentiful than ones aimed at social scientists; as you've already noticed, database usage isn't exactly high in the area. If you were to write a guide yourself on how databases might be useful to social science researchers, I'll bet you'd be one of the first people to write on the subject.
posted by chrominance at 10:49 AM on December 24, 2005

I think it's ultimately an application that enables bad practices, but in this situation, Access is probably the best tool out there. If you can get them to wrap their heads around table relationships, it pretty much walks you through the rest.
posted by mkultra at 11:10 AM on December 24, 2005

Oh, and I don't think web designer tutorials would be particularly helpful here. They're generally geared toward displaying dynamic content, and I think jacobm still needs to output straightforward analysis, where the design of the db matters much more.
posted by mkultra at 11:12 AM on December 24, 2005

they see the very strong advantages databases have over spreadsheets for their needs

Aside:

As mostly non-techy social scientist, what are these advantages? Over, say, a big flat file in Stata?

I might be interested in edumacating myself, but only if I'd see some serious benefit.
posted by ROU_Xenophobe at 11:34 AM on December 24, 2005

Not specifically aimed at social scientists, but a general overview book which appears to avoid heavy tech content is Databases Demystified by Andrew Oppel.

The book has garnered 17 straight 5-star reviews on Amazon, and not all the reviewers could be ringers. Pretty damn impressive actually, considering the savaging a few books I think are great get from a random few of the reviewing masses. No visible carping about it on links outside Amazon, either.

Two possible bonuses are a) it's available as a digital download for immediate perusal and b) the author wrote an (apparently different, likely similar, unreviewed) book called SQL Demystified, if a more SQL-targeted approach is desired. A caveat is that there is disagreement within the reviews over whether the book is for beginners or nontechs: some say yes, some say no.

No personal experience with the book, so it might be a dud for your purposes.
posted by mdevore at 12:54 PM on December 24, 2005 [1 favorite]