The three processes: phonics (a letter by letter sounding out of words); contextual clues (earlier parts of sentences that help readers anticipate upcoming words); and holistic word recognition, or the physical shape of words.
"It's obvious that people must use all three kinds of information to read," says Denis Pelli, an N.Y.U. professor of psychology and neuroscience and senior author of the study published this week in PLoS ONE. "It was far from obvious that you can turn on and off each of the three kinds of information and that you can attribute a percentage of the reading rate to each."
Using passages from author Mary Higgins Clark's murder mystery Loves Music, Loves to Dance, Pelli and study co-author, undergraduate Katharine Tillman, manipulated passages to block readers from using each of the word-deciphering processes.
To muffle context clues, they shuffled words in a sentence ("contribute others. The of Reading measured"); discrimination via word shape was covered up by inserting random capital letters ("ThIS tExT AlTeRnAtEs iN CaSe."); and to eliminate letter by letter decoding, they substituted similar-looking letters into a word, thereby retaining the ability to use word shape and context, once a reader figured out a previous word ("Tbis sartcrec bes lctfan suhsfitufas").
In an effort to determine the impact of the absence of each aid, researchers directed 11 subjects to read passages that were either pure text, had one alteration or two of these manipulations combined. Their findings: that phonics, not surprisingly, the largest component of reading speed, determined 62 percent of the rate. The stunner was that the other two processes consistently contributed the same amounts to reading speed. (Context clues controlled 22 percent of reading speed and word shape governed 16 percent, according to the study.)
Pelli says this research paves the way toward better understanding and addressing the deficiencies of young students who fall behind early in reading.
"It's possible that doing the phonics intervention is the ideal intervention," he says of the most common type of reading training. But "it would be interesting to know how phonics is affecting the other two processes rather than just overall reading speed. We think that it would be very interesting to break down reading rate into these three components in the evaluations of [possible] interventions."