Readability question? (take II)

[Geoff Hart]

Milan Davidovic responded to my cavalier dismissal of readability  
formulas: <<I've seen no readability formula that claimed to work on  
gibberish, nor worked with any technical writer who produced  
gibberish. I therefore wouldn't trust such a test anymore than I'd  
try judging the quality of a baseball bat by playing golf with it.>>

And that's precisely the problem: asking a readability formula to  
tell us whether text is readable is no more logical than playing golf  
with a baseball bat. The tool is not designed to perform this function.

More to the point, the formulas ignore factors that are demonstrably  
far more important than anything the formulas measure: familiarity of  
the words to the audience, grammar, correct punctuation,  
completeness, logic, precision of word use and description, and  
correctness. If any of these factors is inadequate, then it doesn't  
matter whether the readability formula suggests a Grade 1 level or a  
Grade 100 level.

<<The thing to look at is the definition of readability inherent in  
the formula. What factors does it take into consideration, and what  
does it leave out? You then have to decide whether the formula does  
the job for your texts and context.>>

All commonly used readability formulas that I'm familiar with are  
based on various mechanical calculations of things like word length  
and sentence length. For these parameters to be meaningful, you must  
prove that two separate assertions are reasonable: that longer words  
are always more difficult to understand than shorter words, and that  
longer sentences are always harder to understand than shorter sentences.

Neither statement is defensible, even if you replace "always" with  
"most of the time" or even with "more than half of the time". The  
difficulty of a word depends only on whether the word is familiar to  
the reader and whether it's the right word for the job. The  
difficulty of a sentence depends primarily one how well the clauses  
are arranged to support each other and on the other factors listed  
above; length is not relevant if the sentence is well constructed.

If both premises behind the formula are incorrect, then the formula  
can only be correct by random chance, not by design. Q.E.D.: useless!

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --

Geoff Hart   ghart at videotron.ca

(try geoffhart at mac.com if you don't get a reply)

www.geoff-hart.com

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -