The Flesch-Kincaid Test is Flawed and We Should Stop Using It

A while back, we wrote about readability scores in legal blogging and online content marketing. The gist of that post was that there was no reliable evidence to suggest that writing content that was easy to read would do better for search engines optimization (SEO). We recommended that lawyers keep their targeted audience in mind, rather than try to write on a 6th grade reading level all of the time.

One thing that we didn’t delve into was the inherent unreliability of those readability metrics. Let’s do that, now.

Readability Scores

First, we’re talking about readability scores. For online content purposes, these are websites where you can copy your content and paste them into a field and get told all about the mistakes you’ve made. Some examples are:

Some of these tools are more thorough than others, but generally they will give you stats about your content, like:

  • Word count
  • Use of the passive voice
  • How many sentences have more than a certain number of words
  • How many words have more than a certain number of letters or syllables

Generally, the end result is either a grade level or a numerical readability score (higher is easier to read), both of which are based on the Flesch-Kincaid test.

This test, though, is inherently unreliable.

Why is the Flesch-Kincaid Test Unreliable?

It’s fairly simple, really: The Flesch-Kincaid test is unreliable because it uses 2 metrics to grade written content:

  1. Words per sentence, and
  2. Syllables per word.

But the difficulty of the English language doesn’t follow these metrics. Lots of advanced, college-level words have one or two syllables. Plenty of basic words have four or more.

Take, for example, the following wonderful words that are described as “college level vocabulary”:

  • Amok
  • Banal
  • Crude
  • Digress
  • Eschew
  • Forte
  • Haughty
  • Leer
  • Pristine
  • Prone
  • Quibble
  • Sage (someone wise, not the herb)
  • Shrewd
  • Stoic
  • Zealot

And then some 6th grade words that everyone knows, but that have more syllables in them:

  • Anticipate
  • Catastrophe
  • Consequence
  • Democracy
  • Elaborate
  • Ferocious
  • Hypothesis
  • Manipulate
  • Obedient
  • Peculiar
  • Repetition
  • Unanimous
  • Variable

Readable gives the phrase “The haughty zealot observed that the hovel was pristine” a readability score of 75.5 and says that it’s written for 5th graders, even though it has all of these college-level words. But then a sentence filled with 6th grade vocabulary, “Unanimous obedience in a democracy leads to peculiar consequences” is scored a -37.3 (yes, negative points!) and a grade level of 20.7 (is that… post-doctoral?).

There are countless sentences like these that break the Flesch-Kincaid test. It’s time to stop using it to grade online content. The difficulty of an English word does not depend entirely on the number of syllables in it. We all know what a refrigerator is. Few of us know what legion means, when it is used as an adjective.

book with low reading level