Why a grade-level score isn't the same as a reading level

Every popular readability formula spits out a single number labeled "grade level." Flesch-Kincaid says your text reads at grade 9.4. SMOG says 11.2. Gunning Fog says 13. The number looks definitive. Most writers treat it as if it tells them which audience the text is appropriate for. That interpretation is wrong, and the gap between what writers think the score means and what it actually measures explains why "I edited it to a 7th-grade level" articles still feel hard to read.

This article unpacks the actual definition, the historical reason it diverged from "reading level," and the practical implication: what you should be optimizing for instead.

What "grade level" technically means

Every readability formula was calibrated by giving texts to readers and measuring comprehension. The formula then back-solves for "what grade of student would need to read this passage with X% comprehension?"

The X% varies dramatically by formula:

• Flesch-Kincaid Grade Level: 75% comprehension. Anchored to U.S. Navy training-manual studies from 1975.
• SMOG: 100% comprehension. Anchored to McLaughlin's 1969 cloze-test study on health-information texts.
• Gunning Fog: ~90% comprehension. Robert Gunning's 1952 figures.
• Coleman-Liau, ARI: 75% comprehension. Anchored to character-count regressions from the 1970s.

Notice what's not in any of these definitions: the actual reading age of the audience, the audience's vocabulary level, or whether the audience cares about the topic. The grade number is a regression output from a 50-year-old comprehension experiment, not a current-population literacy benchmark.

Why it's not "reading level"

Two reasons this matters in practice.

First, the U.S. average adult reads somewhere between 7th and 9th grade. Roughly 130 million adults are below the level needed to read the average newspaper. If you write at a 12th-grade level and tell yourself "that's college-educated, no problem," you've still excluded somewhere between 50% and 70% of the U.S. adult population. The grade number doesn't tell you that; the population data does.

Second, the formulas only measure surface features. Sentence length, syllable count, character count. They don't measure:

• Whether your nouns are concrete or abstract
• Whether your verbs are active or passive
• Whether the logical flow is signposted
• Whether the reader has the background knowledge required

A text scoring "grade 7" can still be incomprehensible if it's full of unexplained jargon, abstract verbs, and assumes domain knowledge the reader doesn't have. A text scoring "grade 14" can be a breeze if it's well-organized, uses concrete examples, and signals its structure clearly.

The historical drift

The original Flesch and Dale-Chall formulas in the 1940s were designed to help editors at Reader's Digest and the U.S. Department of the Interior write for a specific audience: adults with roughly an 8th-grade education. The grade number was a proxy for "people with this much schooling can read this."

Three things happened to break that assumption:

1. Average U.S. adult education climbed from ~9 years (1940) to ~13 years (2020). The formulas didn't recalibrate.
2. Modern reading habits changed. People skim more, read on smaller screens, juggle more distractions. Text that scored "easy" in 1950 reads harder today.
3. The formulas were never localized. A "grade 9" Flesch-Kincaid in English doesn't translate to a grade 9 reading level in Spanish, Mandarin, or Arabic — different languages have different syllable distributions and sentence structures.

The number persists because it's easy to compute and looks authoritative. The interpretation lags badly.

What to optimize for instead

If you can't trust the grade number to mean "my audience can read this," what can you trust?

Aim for a Flesch Reading Ease of 60+ for general adult audiences. That's "Standard / Plain English" on Flesch's scale, and it's a more population-anchored target than any grade-level number. Most newspapers hit 60-70. Most government plain-language guides target 60+.

Compare formulas to spot specific problems. When SMOG runs noticeably higher than Flesch-Kincaid, you have too many 3+ syllable words. When Coleman-Liau runs higher than both, your character density is the problem (long words, lots of capitalization). When all three agree at a high number, the whole text is dense.

Read it aloud to a real person. Every formula in the world will tell you a 30-word sentence is hard. None of them will tell you whether your reader will care enough to finish the paragraph. The grade number is a hygiene check, not a verdict.

A practical workflow

1. Score your text with the six-formula calculator. Look at the spread between formulas, not just one number.
2. Aim for Flesch Reading Ease 60+ (Standard) or 70+ (Easy) depending on audience.
3. Use the diagnostic guide to identify which specific edit will move the formulas.
4. Read it aloud — out loud, not in your head.
5. Show a draft to one real reader before publishing. Their stumble points beat any formula.

The grade number is useful as a smell test, not a target. Optimize for being read, not for hitting a number on a 50-year-old regression chart.