What "reading level" really means, and why one number misleads

A "reading level" is an estimate of how hard a text is, usually from sentence length and how many uncommon words it uses. It's useful for finding books in the right neighbourhood, and misleading when treated as a precise score. Peabody (peabody.kids) measures every story it writes in code, several ways, and shows grown-ups a level and a range in plain words, not a single number. It's free through the end of 2026.

If you've ever been told your child "reads at a 2.4" and wondered what that means, this post is for you.

Two different things called "reading level"

People use the phrase for two separate things:

1. The level of a text: how hard a particular book or passage is.
2. The level of a reader: which texts a particular child can read comfortably.

Matching the two is the whole game. A child reading text at the right level is practising; a child reading text that's far too hard is mostly struggling, and one reading text far too easy is mostly coasting.

Peabody measures the first one carefully. The second it treats as a starting point that a child's own reading keeps correcting.

How text levels are measured

Most readability formulas look at two things:

• Sentence length. Longer sentences are, on average, harder to hold in your head.
• Word difficulty. Measured either by syllables per word, or by how many words are missing from a list of familiar words.

Here are the six published formulas Peabody calculates for every story:

Each one is marked as valid only where it was designed to be. Peabody never averages them into a "consensus" number, because averaging formulas that were built for different ages just mixes their errors together.

Why one number misleads

Here's the honest finding from the research behind Peabody's levels: readability formulas sort texts into grade bands reasonably well, but on one short story they wander by about a grade. Change a few sentences and the number moves. Two formulas can disagree by more than that.

So a story labelled "2.4" isn't meaningfully different from one labelled "2.9". Treating that decimal as precise leads to bad decisions, like holding a child back from a book they'd enjoy, or pushing them into one that's too hard.

How Peabody measures each story

Peabody's own level for a story comes from two dimensions:

• Sentences, using Flesch–Kincaid.
• Vocabulary, using the share of words outside the Dale–Chall familiar-word list, set against Peabody's own library.

It also works out a range by resampling the story's sentences, which shows how much the estimate could move. The six formulas above sit beside it for anyone who wants to look.

For the youngest readers, Pre-K to 1st grade, formulas don't work well at all. So Peabody measures what matters at that stage instead: words per sentence, words per page, sight words, whether the words can be sounded out at the child's phonics stage, and repetition.

All of this is plain code. It costs nothing to run and gives the same answer every time.

What grown-ups see

With a free grown-up account, Peabody describes a story in words, like "Written for a grade-3 reader: a comfortable read," with the measures explained in plain terms. For comparison only, it shows the published Lexile and guided-reading ranges usually given for that grade, with a clear note that Peabody measures neither of them.

Children see none of this. A child never sees a score or a level number about themselves; on the kid side, books just sit on floors of the library tower.

How Peabody finds a child's starting level

There's no test. The first time a child makes a story, they tap the sentence they can read comfortably from a handful that run from "I can see you." to "Above the timberline, the wind turned cold and the clouds began to build." That's their starting level, and they can change it with the 📏 button.

When the books a child reads look consistently too easy or too hard, Peabody can offer "Let's find your books": five quick taps at most, with five dots showing progress and no score at the end. In testing on simulated readers, it lands within one level of the right answer about 85% of the time, and a child's later reading corrects the rest.

Lexile, guided reading and grade levels

Schools often use other systems:

• Lexile measures, a commercial scale from MetaMetrics.
• Guided reading levels (letters like F or M), assigned by teachers using a book's features and a child's reading.
• Grade-level equivalents from standardized tests.

Peabody doesn't produce any of these, and its levels aren't a substitute for them. If your child's school uses one, the most useful thing is to ask the teacher what range they suggest, then try a Peabody story a floor or two either side and see how it feels.

Try it tonight

Pick two stories from the free library, one floor apart. Have your child read the first page of each to you. The one that sounds smooth, with one or two stumbles at most, is the right neighbourhood. That's a more useful reading level than any decimal.

Questions parents ask

What reading level is my child?

A rough guide: the level where they can read a page aloud smoothly, stumbling on only a word or two, and tell you what happened. Peabody's library is sorted from Pre-K to 8th grade so you can try a floor and adjust. For a formal assessment, ask your child's teacher.

Does Peabody give my child a reading level score?

No. Children never see a score. Grown-ups see a level described in words, with a range, not a precise number.

Is a Peabody level the same as a Lexile level?

No. Peabody doesn't measure Lexile. For comparison only, grown-ups see the published ranges usually given for a grade.

Why do different websites give different levels for the same book?

Because they use different formulas, built for different ages, and on short texts those formulas can disagree by a grade or more. Treat any single number as a neighbourhood, not an address.