Readability scores are meant to answer a straightforward question: how difficult is a piece of writing to read?
For ordinary prose, that question can often be approached by looking at things such as sentence length, word length, syllable count, and the proportion of complex words. Longer sentences and longer words generally demand more from a reader, so formulas built around those features can provide a useful estimate of reading difficulty.
Poetry complicates nearly every assumption behind those formulas. A poem may contain only one grammatical sentence spread across several stanzas. It may use fragments intentionally. Its line breaks may slow down a phrase that would otherwise be simple, or make a familiar sentence look far more complex than it is. A poem may rely on short, common words while remaining difficult because of syntax, metaphor, historical distance, compression, or ambiguity. It may also use long words in a context that is perfectly clear.
Traditional readability formulas are not useless when applied to poetry, but they need to be kept in perspective. They were largely designed to evaluate prose, and poetry does not always behave like prose. To see how dramatically these measures can disagree, I compared three familiar poems: William Carlos Williams’s “The Red Wheelbarrow,” Emily Dickinson’s “Hope is the thing with feathers,” and William Shakespeare’s Sonnet 3.
Three Poems, Very Different Results
Here are the traditional readability scores produced for each poem:
| Measure | The Red Wheelbarrow | Hope is the thing with feathers | Sonnet 3 |
|---|---|---|---|
| Flesch Reading Ease (%/100) | 74.270 | 35.035 | 76.820 |
| Flesch-Kincaid Grade (Grade Level) | 6.875 | 15.680 | 25.514 |
| Gunning Fog Index (GL) | 6.400 | 17.752 | 28.180 |
| Automated Readability Index (GL) | 8.059 | 18.803 | 31.500 |
| Coleman-Liau Index (GL) | 9.177 | 7.505 | 6.780 |
| SMOG Index (GL) | N/A | N/A | N/A |
The disagreement is immediate.
“The Red Wheelbarrow” receives relatively accessible scores across most measures, which is unsurprising. It is short, uses familiar vocabulary, and contains few long words.
Dickinson’s poem is treated very differently. Flesch Reading Ease gives it a score of 35.035, while Flesch-Kincaid places it at grade 15.680. Gunning Fog and the Automated Readability Index push it even higher. According to those formulas, the poem looks more like advanced academic writing than a lyric built largely from familiar words and a clear central metaphor.
Sonnet 3 produces the strangest pattern. Flesch Reading Ease rates it at 76.820, slightly easier than “The Red Wheelbarrow.” At the same time, Flesch-Kincaid places it at grade 25.514, Gunning Fog at 28.180, and the Automated Readability Index at 31.500. Coleman-Liau, meanwhile, gives it the lowest grade estimate of the three poems at 6.780.
The same poem is being described as highly readable, graduate-level, and roughly middle-school level, depending on which formula we consult. That does not mean the calculations are necessarily broken. It means they are measuring different things, and poetry exposes the differences between them.
What the Traditional Measures Are Looking For
Flesch Reading Ease and Flesch-Kincaid Grade both rely heavily on sentence length and syllables per word. Flesch Reading Ease converts those elements into a scale where higher scores indicate easier text. Flesch-Kincaid uses similar inputs but translates the result into an estimated grade level. These formulas work reasonably well when sentence boundaries behave predictably. Poetry often disrupts that condition. A poem may contain a long grammatical sentence divided into short, manageable lines. The formula still sees the sentence as long, even if a reader experiences it in smaller units.
This helps explain Dickinson’s result. “Hope is the thing with feathers” is not written as a series of short prose sentences. Its punctuation and stanza structure are poetic, not expository. A prose-oriented formula can interpret that structure as sentence-level complexity even when the vocabulary itself remains approachable.
Gunning Fog also uses sentence length, but adds the proportion of complex words, usually defined through syllable count. It attempts to estimate the years of formal education a reader might need to understand the text. That approach can become unstable in a short poem because a small number of longer words can have an outsized effect.
The Automated Readability Index uses characters per word and words per sentence rather than syllables. This makes it somewhat easier to calculate automatically, but it preserves the same basic assumption that longer words and longer sentences generally increase difficulty. Again, that assumption is not unreasonable for prose. In poetry, however, character length may not reflect actual familiarity. An older word may be short but unfamiliar. A longer word may be common and easily understood. A Shakespearean sonnet can contain historically distant phrasing and inverted syntax while still using many short words.
Coleman-Liau also works with characters rather than syllables, but it uses a different relationship between letters, words, and sentences. In this comparison, it produces the opposite ranking from several of the other formulas. It gives Dickinson a lower grade estimate than Williams and gives Shakespeare the lowest estimate of all. That result is useful because it shows how differently a poem can look when sentence length and letter density are weighted in another way.
SMOG (Simple Measure of Gobbledygook) takes yet another approach. It focuses strongly on polysyllabic words and is generally intended for longer samples containing enough sentences to produce a stable estimate. None of these poems provides a sufficient sample, so the index returns no score. That absence is not a flaw in the analysis. It is an honest indication that the texts fall outside the conditions the measure expects. Very short lyric poems simply do not provide enough material for every prose readability formula to operate reliably.
Why Poetry Creates Problems for These Formulas
The central issue is not that poetry ignores language. It is that poetic difficulty is distributed differently from prose difficulty. Sentence length is one example. A poem may have a long sentence but short lines, clear syntax, and familiar vocabulary. Another may contain several short sentences that are difficult because of fragmentation or ambiguity. A formula that focuses primarily on sentence boundaries may mistake punctuation for cognitive burden.
Word length creates similar problems. Long words are sometimes more difficult, but not always. Short archaic words, unusual spellings, dialect forms, and syntactic inversions may be far more challenging than a familiar three-syllable word. Character count and syllable count do not measure historical familiarity.
Poetry also uses lineation as part of meaning. A line break can create emphasis, suspense, rhythm, double meaning, or visual structure. Traditional readability formulas usually do not account for the fact that readers process poems line by line as well as sentence by sentence.
Then there is interpretive complexity. A poem can use easy words and still be hard to understand. “The Red Wheelbarrow” is a perfect example. Almost every word is familiar, but the poem does not explain what “depends” upon the scene it presents. Its linguistic surface is accessible, while its meaning remains open. The opposite can also be true. A poem may use older or less familiar language while presenting a relatively straightforward argument. Readability formulas often merge these forms of difficulty into a single number, even though they are not the same experience.
Introducing VerseVAD Poetry Reading Ease
The VerseVAD Poetry Reading Ease score, or VV-PRE, was designed specifically to estimate the linguistic accessibility of poetry.
Instead of relying primarily on prose sentence length, VV-PRE combines four components that are especially relevant to poems:
- word frequency;
- age of acquisition;
- line accessibility;
- word complexity.
Word frequency asks how common the poem’s content words are in general language use. A poem filled with familiar vocabulary will usually be easier to process than one built from rare or specialized terms.
Age of acquisition estimates how early readers typically learn the poem’s words. Early-acquired vocabulary tends to feel more immediately available than words learned later in life.
Line accessibility looks at the amount of lexical material placed on each nonblank line. This gives lineation a role in the score rather than treating the poem only as prose with unusual spacing.
Word complexity considers the average number of syllables in the poem’s content words. Longer words still matter, but they are not allowed to determine the entire result. The resulting scores for the three poems are:
| Poem | VV-PRE (%/100) | Classification |
|---|---|---|
| The Red Wheelbarrow | 80.974 | Accessible |
| Hope is the thing with feathers | 71.544 | Accessible |
| Sonnet 3 | 62.046 | Moderately Demanding |
This ordering feels much closer to the reading experience these poems are likely to create.
“The Red Wheelbarrow” receives the highest score. Its words are familiar, its lines are extremely short, and its vocabulary is generally acquired early.
Dickinson’s poem remains accessible, but it scores lower. It uses somewhat more complex phrasing and a larger lexical range than Williams, even though it is still far more approachable than its Flesch-Kincaid grade of 15.680 or Gunning Fog score of 17.752 would suggest. These scores would suggest that the near completion of a bachelor’s degree is needed to access one of Dickinson’s most popular poems.
Sonnet 3 falls into the Moderately Demanding range. Its syntax, vocabulary, and historical language place more pressure on the reader than the other two poems. VV-PRE recognizes that difference without claiming that the sonnet requires thirty-one years of education, as its Automated Readability Index might appear to suggest.
What VV-PRE Does Not Measure
VV-PRE is not intended to measure a poem’s quality, intelligence, beauty, originality, or interpretive depth. A higher score does not make one poem better than another. It does not mean that a poem is appropriate only for younger readers, nor does a lower score mean that a poem is more sophisticated. The score is about linguistic accessibility.
“The Red Wheelbarrow” can be easy to read and difficult to interpret. Dickinson can use familiar language while still creating metaphorical and philosophical complexity. Shakespeare can be linguistically demanding without being incomprehensible. VV-PRE is designed to preserve those distinctions.
It also does not make traditional readability measures unnecessary. Flesch, Gunning Fog, ARI, Coleman-Liau, and SMOG can still provide useful information about sentence structure, syllable patterns, word length, and the limits of particular formulas. The goal is not to replace every existing score with one new number. The goal is to keep the numbers in perspective.
A Better Basis for Comparing Poems
The most revealing part of this comparison is not that the formulas disagree. It is why they disagree. Traditional readability measures see Dickinson’s long poetic sentence, Shakespeare’s compressed syntax, and Williams’s short words through the assumptions of prose. Each formula emphasizes a different combination of sentence length, syllables, letters, and complex words. Poetry pushes those assumptions in unusual directions, so the results can become contradictory.
VV-PRE approaches the question from another angle. It asks whether the poem uses familiar words, when those words are typically acquired, how much lexical material appears on each line, and how complex the content words are. That does not solve every problem involved in measuring poetry. No formula can fully account for metaphor, irony, historical context, ambiguity, rhythm, or the experience of an individual reader. It does, however, give us a more appropriate way to compare linguistic accessibility across poems.
Readability scores are most useful when they help us ask clearer questions. Why does one poem feel more immediately approachable than another? Is the difficulty coming from vocabulary, line structure, word complexity, or syntax? Does the poem use simple language to create complex meaning? Traditional measures give us part of that picture. VV-PRE adds a poetry-sensitive perspective. The result is not a final judgment about how difficult a poem is. It is a more grounded place to begin.