How reliable/relevant are readability scores?

a) All of them. After all, they do measure something.
b) You could improve grammar and vocabulary. But it's not really necessary. "Winnie the Pooh" is a very easy to read book enjoyed by all ages. Content is important, too.

Readability scores are pretty simplistic things, and the grade levels aren't validated at all. They're merely rough guides. If you want to game your score you can do it by using more polysyllabic words and longer sentences. These aren't the only things making a sentence more or less readable. There's also prosody, word choice, word order, vocabulary ... what's easier to read, "the elephant is relevant" (Flesch-Kincaid readability level 9.6), or "the sixth sick sheik's sixth sheep is sick" (Flesch-Kincaid readability level -0.7)? The first of course, but their Flesch Kincaid scores suggest the opposite. I'm not sure this test is completely useless, but it's getting there.

P.S. I ran my own post through a readability checker. It gave it a grade level of 8.1. I'm insulted. I thought it was completely​ unreadable.

Albedo said:

... what's easier to read, "the elephant is relevant" (Flesch-Kincaid readability level 9.6), or "the sixth sick sheik's sixth sheep is sick" (Flesch-Kincaid readability level -0.7)? The first of course

Click to expand...

In case you were wondering: the first sentence is more difficult to comprehend. It requires additional thinking to understand what exactly it means by "relevant". It's way too "meta". While the second sentence is very easy to understand since everybody can easily imagine what it means for a sheep to be sick. The tongue-twister quality of the text matters only if you read it aloud and that's not what reading tests are about.

I don't know anyone in publishing who uses scores like those. They just read the text and see if they like it.

Educational publishers use these to level books and passages for things like early readers and comprehension tests. They don't have much of a place in fiction otherwise. They use metrics for number of syllables in each word, words per sentence, and the like. (However, they can help you see how your characters sound. So more erudite and wordy characters will have a higher level than a four year old. Of course, the character could appear erudite but make no sense.) Stay true to your voices and ages and the level should be where you need it. Most eighth graders don't speak at this level, they merely read at it. They'd get laughed at if they used eighth grade vocab in their daily chitchat except for effect.

Readability scores are crude but useful tools. As long as you keep their weaknesses in mind.

Grade level, for instance, in many of them means longer words and sentences. It doesn't mean your writing is "smarter" but that it's dumber. Convoluted writing makes it harder for even the smartest of us to get past the words to the meaning beyond them. A lower grade level is (on the average) good.

Some of the immortal and most profound writing is deceptively simple.

What I did to help my writing style was to create a tool that turns the background of all three-syllable words light grey, four light yellow, and five or higher red. When I've finished a chapter I go through each red word and see if its length is justified. If not, I seek a more expressive word. Ditto the yellow words.

I give only a cursory glance at the light grey words, but I do consider them for an instant.

Curlz said:

In case you were wondering: the first sentence is more difficult to comprehend. It requires additional thinking to understand what exactly it means by "relevant". It's way too "meta". While the second sentence is very easy to understand since everybody can easily imagine what it means for a sheep to be sick. The tongue-twister quality of the text matters only if you read it aloud and that's not what reading tests are about.

Click to expand...

I struggled to read the second sentence. All the words looked the same. It didn't make sense, and I had to put in some serious work to decipher it. First sentence? Not a problem, although it's a cliffhanger. Now I want to know what the elephant is relevant to.

Old Hack said:

I don't know anyone in publishing who uses scores like those. They just read the text and see if they like it.

Click to expand...

They've not been respected in terms of most k-12 instruction for a very long time. They're based on old data, and they've not been updated.

Curlz said:

In case you were wondering: the first sentence is more difficult to comprehend. It requires additional thinking to understand what exactly it means by "relevant". It's way too "meta". While the second sentence is very easy to understand since everybody can easily imagine what it means for a sheep to be sick. The tongue-twister quality of the text matters only if you read it aloud and that's not what reading tests are about.

Click to expand...

I'd argue that a primary school child is much likelier to know what an elephant (or a hippopotamus, or a muttaburrasaurus) is than what a sheik is, though. Scores based on word length alone ignore that the number of syllables in a word isn't necessarily an indicator of the word's complexity or rareness.

A lot of readers, especially children, do read aloud, and even many adults move the vocal tract silently when concentrating on a passage. So elements of prosody, assonance, articulation etc. do matter when reading.

Anyway, my point is that the most common readability test gives wildly different scores to those sentences, evidently just because the first has two 3-syllable words, and the last is made up entirely of monosyllabic words. I don't think it can be said to be a reliable measure of readability.

AW Admin said:

They've not been respected in terms of most k-12 instruction for a very long time. They're based on old data, and they've not been updated.

Click to expand...

Lexile updated with the common core (I think in 2012). A further update occurred last year: https://thejournal.com/articles/201...xile-framework-for-students-in-grades-k2.aspx I have been asked by packagers and educational publisher to write to specific Lexile levels. To level a text, I do remove scientific terms and proper nouns and replace them with pronouns. Those words are necessary to the text but they skew the level. (For example: "Juan Pablo Montoya won the Indianapolis 500" is much harder to read than "He won the Indy 500.")

Keep in mind also that these books often contain glossaries and images to help with comprehension. A picture of a hippo will make the word far clearer than a picture of a sheik does. Sheik needs a definition.

It might also be good to note that many newspapers are written at a fifth grade level. (At least that's the last number I heard.)