Reading about Readability

Photo: On the platform reading. (cc) Mo Riza.

Bora Zivkovic highlighted a post on readability and blogging yesterday on The Science of Blogging. This is a rework of a post on Colin Schultz’s site from earlier in the month, which is well worth a visit because of the comments on the post.

Colin Schultz shows how you can use Google to get a readability score for your blog. The score is between 1, the simplest and 3, the most difficult. He also shows some scores for popular science blogs. The median is 2 and AoB Blog comes in at 2.55. Very few blogs score higher, and it puts this site in the top 10% for difficulty. Obviously as I’m not happy with this score I shall now proceed to rubbish the method by mocking a system that marks marmalade as a difficult word.

Actually there’s no need. Schultz is well aware of the limitations of readability tests. In fact the simplicity of his test is that it divides posts into three categories rather than a spuriously precise score and derives a ball park figure from that. It might not be wholly accurate, but it’s good enough to show that a blog post that uses words like spuriously and derives is going to be difficult. So instead I’ve been thinking about what the test means.

One reason the blog scores so high is that the bulk of entries are ContentSnapshots. These are summaries of the papers we have in the journal. They’re easier to read than abstracts, but they’re still difficult. Should I use the scores to ask Pat to make them simpler? I don’t think so. Their purpose is to let you know what the papers are about. If you can’t follow the snapshots, then you’re not likely to be able to read the papers. It is at a high reading level, but that’s the target audience for them. I was going to make this point, that blogs mean you can cater for a variety of reading skills, a Big Idea. Bora Zivkovic has already beaten me to that, including how blogs can create a richer reading ecosystem by deliberately cross-linking between levels.

Schultz also makes the very reasonable point that quick throwaway posts could lower a score. Maybe the natural level for a Science Blog is around 2. I’m wary of that though. It suggests that one way to make your writing more sciencey is to make it more difficult to read and that’s like a cargo-cult idea of science. I’ve seen places where it happens, but it’s not something we should encourage. Ed Yong in the same comments thread notes his readability score has gone down. His writing might have become more accessible, but it’s certainly not dumbed-down.

I think an Einstein quote is useful here. Everything should be made as simple as possible, but not one bit simpler. I think that there will always be some complexity in scientific language. Discovering new concepts means we will need new words. At the same time complexity isn’t inherently a virtue, and I think Schultz’s method is a quick and easy way to help make things as easy as possible.