Monday, June 14, 2010

Readability Statistics: Are They Meaningful?

I am a recovering scientist. In my prior life, my writing was copious, erudite and replete with $50 words. Copious, erudite and replete being examples. I even won awards for writing arcane trivia with a lavish lexicon. As I moved into literary nonfiction, however, a senior editor advised me to cut the crap. She told me I would be publishable only if I wrote at the 7-8th grade comprehension level and got my characters per word down below 4.5. A journalism professor cautioned me about the SMOG Index (Simple Measure of Gobbledygook) which penalizes words of three syllables or more. But then a noted nonfiction writer, a professor of English, told me in rather pithy terms that readability stats would not make me a good writer, that they were a distraction, and that I was wasting my time. I felt caught in the middle.

To find some answers, I analyzed several 1000 word writing samples from each of the godparents of nonfiction, those writers I strove to emulate. On the list were Annie Dillard, Barry Lopez, Barbara Kingsolver, Joan Didion, Terry Tempest Williams, and others. I also sampled more than 50 others for comparison, including contest winners, essays in selected literary journals, lesser known book authors, and the like.

“And what to my wondering eyes did appear” but a really strong, narrowly defined pattern. Averages calculated for each of three groups, nonfiction godparents, edited literary journals, and contest winners, all fit these parameters. There were almost no exceptions. The editor had been right. Characters per word averaged below 4.5 in all cases. Sentences were short, punchy, averaging between 15.1 and 16.6 words. Measures of Flesch-Kincaid Reading Ease fell between 72.4 and 75.6, which, within a range of 1-100, is basically homogenous. Average grade level for full comprehension measured between 7.7 and 8.0, again nearly identical. Average SMOG indices for each group were either 6.7 or 6.8. And last but not least, passive voice occurred in less than 5 percent of all sentences written by these authors. Parenthetically, the lesser known authors violated every one of these parameters, always leaning in the direction of worse writing.

My English professor friend was also right. Readability statistics cannot turn anyone into a great writer. Soul and substance are key, and the art and craft that captures them in a latticework of words makes for greatness. But, like chalk lines on a sports field, readability statistics loosely set the boundaries within which the game is played. And, like it or not, many editors and agents have come to expect submissions to fall within these same gridlines.

What are your own readability stats? Compare a first draft with a finished product.


  1. I have never had a problem filling my writing with $50 dollar words-I'm more of a $19.99 kind of gal. You know I have never been big on focusing on this readability stuff when I am writing--I just write. But in the name of science, I checked an article I am getting ready to send off tomorrow and for the most part it falls in the averages you were talking about (without me even knowing!). So I am right there with the likes of Kingsolver, Williams and Dillard. Woo Hoo!

  2. $19.99, huh? That sounds about ideal. But then that is just my $0.02 worth. Readability stats' main use may be to apply light wrist slaps for major transgressions. I find the biggest violations, at least in papers read in graduate classes and critique groups, are passive voice at 7-8 percent or higher, and sentence length at 22-23 words or longer. More than 25 words and you tread, almost perforce, into the arena of compound and complex sentences, replete (love that word) with an overdose of prepositional phrases. After that, I turn to any measure that captures to overuse of multisyllabicisms. :-)