Plain English isn’t dumbing down: what readability actually measures

A common assumption treats simple sentences as a sign of simple thinking, and long, elaborate ones as evidence of a sharper mind behind them.

The research on how readers actually judge writing points the other way, and the tools built to measure readability were never designed to reward simplicity for its own sake.

They were built to answer a narrower, more useful question: how much effort does this text demand from the person reading it.

What a readability score is actually counting

Readability formulas don’t evaluate ideas, arguments, or style.

They count.

The most widely used version, the Flesch-Kincaid formula, emerged from a 1975 project for the U.S. Navy, where researcher J. Peter Kincaid and colleagues needed a way to match technical training manuals to the reading level of enlisted personnel. Their report, tested against 531 sailors across four training schools, derived a formula from two measurable quantities: average sentence length and average syllables per word.

Feed those two numbers in, and the formula outputs a rough U.S. grade level.

That’s the entire calculation. It says nothing about whether the sentences are well-argued, whether the vocabulary is precise, or whether the piece has anything worth reading in it.

A readability score is a proxy for processing effort, not a verdict on quality, and treating it as the latter is where a lot of the confusion about “dumbing down” comes from. A text can score as extremely readable and still be shallow, and a text can score as demanding and still be excellent.

The formula was built to solve a training-manual problem, not a literary one.

Complex language doesn’t make writers look smarter — it makes them look less smart

Given that readability scores don’t judge quality, it’s worth asking what unnecessarily complex writing actually buys a writer. Daniel Oppenheimer tested this directly, and the result is one of the more cited findings in the psychology of writing.

Across a series of experiments, Oppenheimer’s research found a consistent negative relationship between the complexity of a text’s vocabulary and how intelligent readers judged its author to be — the effect held regardless of whether the underlying essay was actually well argued, and regardless of what readers expected going in. The mechanism behind it was processing fluency, not snobbery running in reverse: text that’s harder to read triggers a felt sense of difficulty, and readers attribute that difficulty to the writer rather than to the words on the page. A separate experiment in the same study found the effect wasn’t limited to vocabulary — texts set in a harder-to-read font were judged to come from less intelligent authors too, even when the words themselves were unchanged.

The practical upshot cuts against the instinct to reach for the more impressive-sounding word. Needless complexity doesn’t read as sophistication. It reads as friction, and readers assign the friction to the writer’s ability rather than to a deliberate choice.

Governments learned the same lesson at a much larger scale

This isn’t only a matter of individual writers making stylistic choices. Entire governments have legislated against needless complexity after watching what it costs in practice. The U.S. Plain Writing Act of 2010, signed into law as Public Law 111-274, requires every executive branch agency to write public-facing documents in language that is, in the Act’s own definition, “clear, concise, well-organized, and follows other best practices appropriate to the subject or field and intended audience.” Agencies have to train staff, publish compliance reports, and apply the standard to anything a citizen needs to read in order to access a government service, understand a benefit, or meet a legal obligation.

The law exists because unclear government writing has a measurable cost: forms that go unfilled correctly, benefits that go unclaimed, instructions that generate support calls and appeals. Plain language, at that scale, is the difference between a document that does its job and one that doesn’t, not a matter of style preference.

What a readability score can’t tell you

None of this means a low grade-level score is a mark of good writing on its own. A formula that only counts sentence length and syllables can be gamed by chopping sentences apart until the prose is choppy and hard to follow for entirely different reasons, or by swapping in short words that are vaguer than the long ones they replace.

Readability metrics say nothing about whether ideas are sequenced logically, whether transitions are doing their job, or whether the piece actually answers the question a reader brought to it. A manuscript can hit a low target grade level and still be confusing, and a manuscript aimed at a specialist audience may reasonably score higher without being badly written.

The actual discipline of plain English is a deliberate choice, made sentence by sentence, about what a reader needs from this text — cutting everything that makes them work harder to get it — rather than a matter of shortening words until a formula is satisfied. That’s a harder skill than reaching for the longer word, not an easier one, and the research on how readers judge complexity suggests they can tell the difference.