Read: Readability formulas are not recommended
Naturally, a writer would like to know whether his text is suitable for the intended target group. It is therefore tempting to use a legibility formula. After all, the score indicates the difficulty of the text. Is that right? Are you considering using a legibility formula, or have you read somewhere in a booklet with writing advice not to write sentences longer than 20 words? Then information about the usability of readability formulas is certainly desirable.
Side note on the use of legibility formulas
Various legibility formulas have been developed to measure the legibility of a text. The result of the test (the score) indicates the difficulty of the text. A low score for Flesch’s readability formula means that we are dealing with a text that is difficult to read, suitable for readers with an academic education. Writers naturally want to know whether their text is legible for the readers for whom the text is primarily intended (the intended target group). The use of a legibility formula therefore seems attractive. After all, the score indicates whether the text is not too difficult for the intended target group. In addition, a score that is too low, indicating that the text is too difficult for the intended target group, can be used to rewrite the text. Cutting long sentences, replacing single commas with periods and using shorter words leads to a lower score, and thus to a more comprehensible text.
It sounds too good to be true. It probably is. A copywriter who wants to use a readability formula should consider whether the score of a readability formula is reliable. Is it possible to measure the readability of a text using the readability formula? To be able to answer that question, we will examine two readability formulas, namely Rudolf Flesch’s readability formula and the cloze test.
Rudolf Flesch’s legibility formula
In 1949, the American writing consultant Rudolf Flesch made his readability formula public. He presented his readability formula as part of a theory on reading and writing problems.
Flesch strives for a personal, informal way of writing. He hates pomposity and abstract discourses. The person-oriented spelling propagated by him is used in the magazine Het Beste (by Readers Digest). In this magazine, an article does not only draw attention to a problem. It also describes how someone solved the problem.
The elements that, according to Flesch, determine the readability of a text are human interest and ease of reading. Human interest is about whether the text is of interest to the reader. A text is interesting for a reader if he can identify with the persons and events in the text read (perhaps the reader can identify with the person who solves a problem in Het Beste). The ease of reading a text is determined by the substantive difficulty of the text, the sentence structure, the word length and the sentence length.
Calculating the reading convenience
Flesch is of the opinion that the ease of reading a text mainly depends on the word and sentence length. A text with short words and sentences is an easy text. After all, research has shown that difficult texts contain on average many long sentences and long words.
For ease of reading, Flesch has developed the following formula:
206.835 – (mean sentence length X 1.015 = number of syllables per 100 words X 0.846)
In order to arrive at a correct result, three to five samples of 100 words must be taken per article. Average sentence length is calculated by dividing the number of words by the number of sentences. Average word length is calculated by counting the average number of syllables per 100 words.
A low score means that the text is difficult to read, while a high score (maximum 100) means that the text is easy to read.
Reliability of the Flesch formula
Does this formula actually measure the reading convenience of texts? That question can be answered by putting it to the test. The following excerpt from the story My secret son comes from Archives Reve 1961-1980, page 24.
Counting the number of sentences and the number of syllables per 100 words, of this excerpt and of some other passages in the text, results in an average sentence length of 47 words and a word length of 161 syllables per 100 words. If we enter these numbers in the readability formula, we arrive at a score of 23, which would mean that it is a difficult text, suitable for academics. Despite the long sentences, the text does not pose any major problems for the reader. It is certainly readable for high school students. The reliability of Flesch’s readability formula can therefore be questioned, because this text is not as difficult to read as the score suggests.
Linguistic variables that influence the ease of reading of a text
There are many variables that influence the ease of reading of a text. Those variables may have more influence on the reading convenience of a text than sentence length and word length. The main variables are:
- The presence of tong constructions (belonging together words in a sentence are far apart, for example: ‘He said, looking out the window, that he couldn’t help him)‘.
- The word frequency. There are frequency tables that indicate whether a word occurs relatively often in Dutch.
- The number of person forms per sentence (more person forms per sentence make a sentence easier to read).
- The number of nouns per sentence (more nouns make a sentence more difficult to read).
- Percentage personal pronouns (me, you, us, and so on) per sentence
The usefulness of Flesch’s formula
The score of Flesch’s formula is, as we have seen with Reve’s text, not reliable (not in accordance with the difficulty of the text). This is because, as we have seen, there are linguistic variables other than word and sentence length that affect the readability of a text. In addition, the readability of a text is also determined by a number of other factors:
- Psychological factors, such as the context (environment in which text is used) of the text and the reader’s background knowledge.
- The way in which sentences are connected. How are relationships expressed, are reference words used and how are the different parts of the text connected?
So there are many linguistic and non-linguistic factors that determine the readability of a text. Long sentences and words are therefore often not the (main) cause of the difficulty of a text. Flesch’s legibility formula is therefore no writing advice. Using short words and only making sentences with a maximum sentence length of 20 words does not lead to easily readable texts. Cutting long sentences usually does not have the desired result (an easier readable text).
Flesch’s readability formula is not very useful because too much weight is given to the linguistic factors sentence length and word length. In addition, Flesch’s readability formula cannot be used as writing advice. Is there a formula that might be more useful?
The cloze test
In this text every umpteenth word (often the fifth word) is omitted. The omitted words must be completed later. The number of correctly entered words provides an indication of the readability of a text. Usually people work with a length of 250 words (so 50 open spaces).
We take the test … the sum. Do you think … this cloze test is more reliable … Flesch’s readability formula? … you don’t have this question yet …..?
The uncertainty is the main objection to the cloze test. Compare the following sentences:
Between all the toys, the …. was invisible.
Which word must be entered in the second sentence? That is difficult to determine. However, that does not make the second sentence more difficult than the first sentence. The reliability of the cloze test can therefore also be questioned.
It is of course important to write legible texts. In my opinion, the use of legibility formulas does not lead to legible texts. That’s why I don’t use them myself. Of course, that doesn’t mean that copywriters shouldn’t use them.
Anyone who delves into Rudolf Flesch’s work will undoubtedly also come across the formula he developed for measuring the human importance of texts. Also for that formula, copywriters can apply that formula if desired.