Cross-posted from Fluff-n-Stuff
——————————————————————————
Googling for something else, I found this link which purports to gauge the readability of your blog. It comes with a nifty piece of code that you can use to embed it in your page as well, and I was highly flattered when it tagged my blog as being in the ‘College Post-Grad’ level. What marvels of prose made it adjudge my blog to be worthy of that honor, I wondered. More importantly, was it my deathless prose, or my ease of using French phrases or poetry quotes?
What was the algorithm behind this rating?
I started checking out all my favorite blogs and came up with a bunch of mixed results: Some that I rated very highly came up with ‘High School’, while others such as ‘rushlimbaugh.com’, ‘microsoft.com’, ‘google.com’ came up with ‘Genius’. Could this be a sort of negative rating , meaning the higher you score, the more meaningless the result?
There are, fortunately, more reputable ways to check on the readability index, as this suite of tests provided by Juicystudio. The disclaimer for the interpretation states:
This service analyses the readability of all rendered content. Unfortunately, this will include navigation items, and other short items of content that do not make up the part of the page that is intended to be the subject of the readability test. These items are likely to skew the results. The difference will be minimal in situations where the copy content is much larger than the navigation items, but documents with little content but lots of navigation items will return results that aren’t correct.
It’s a relief to know that Rush Limbaugh is less likely to be a genius, and just has a lot of navigation items on his page that have skewed the results.
Ruchira adds:
The results are indeed intriguing. Leiter Reports is predictably "genius." I thought A.B. would come out at "college level, undergrad" but was "high school" instead. The kicker is 3 Quarks Daily which is chock-full of art, science and philosophy articles and boasts a stable of authors many of whom are either Ph.Ds or M.Ds, came out at Junior High reading level :-)
I wonder if the assessment is skewed not just by navigational links but also by the number of images displayed on the blog (Leiter – none, 3 QD – lots), the consideration being : more pictures = easier content.
Accidental Blogger 
2 responses to “Reading Level (Sujatha)”
Great post Sujatha. What you and Ruchira show is that computers haven’t quite arisen to the level of human beings in doing textual interpretation — yet. And that’s probably for the best. Who needs an algorithm to tell you that Rush Limbaugh arises to the level of “genius.” Yeah right, he arises to the level of Ulysses or Thomas Hobbes.
Anyhow, if you’re really into this topic, Richard Powers wrote a book about 10 years ago called Galatea 2.2, in which a writer grapples with the fact that a new computer is able to interpret English literature as well as a person. It’s ultimately a meditation on whether computers can be programmed to be capable of humanistic models of evaluation (i.e. beauty, justice, morality), rather than just value-neutral computation. As a book, I found it hard to get into, and never finished it, but I admire it conceptually.
LikeLike
The problem with all these methods is that it is only as good an approximation of real life as the algorithm can handle. Even if the ‘readability’ test had incorporated some kind of fuzzy logic, it would not be able to mimic a human sense of aesthetics. Perhaps in the (not very near) future, the readability test may incorporate rules regarding generally accepted moral and justice principles, in which case it might point out that the Limbaugh website is somewhere in the ‘Moron’ category.
LikeLike