Content analysis usually requires the researcher to establish inter-coder reliabillity, which means that about 10 percent of the articles in a sample frame are content analyzed by two different coders in order to make sure there aren't wildly different interpretations of text. (In theory, this keeps the research honest, although it's probably the case that in the end both coders will have certain shared assumptions about what they're coding.) For the study to be legit, inter-coder reliability should be around 80 percent or higher. For the inter-coder reliability to be legit, though, it must rely on truly randomly selected articles at the outset.

Sometimes this can be a pain. Researchers generally rely on all manner of goofy ways to get a good batch of random numbers, none of which involve computers. What? No computers? That's right, no computers. The problem with computer-generated "random" numbers is that they're not truly random, as any programmer or computer scientist could tell you. These numbers might appear random, but they usually use a complex mathematical formula that, in the end, has predictable results.

It looks like Random.org might have found a way around all this, though. The site uses atmospheric noise (read: the weather) as a starting place for generating randomness. I don't know whether many (or any) researchers have latched onto this, but I don't see any reason in principle why they couldn't. Given that the site has apparently been around since the late 90s, in fact, I'm actually pretty astounded that this method wasn't covered in my Research Methods course last semester.

To make this relevant for my purposes, it looks like I don't have to waste time pulling a hundred different numbers out of a hat to perform my intercoder reliability. How cool is that?

Be first to comment this article

Only registered users can write comments.
Please login or register.

Powered by AkoComment Tweaked Special Edition v.1.4.6
AkoComment © Copyright 2004 by Arthur Konze - www.mamboportal.com
All right reserved