Studies Cast Doubt on Benefits of Using Only Whole Language To Teach Reading
Three newly published studies cast doubt on the effectiveness of whole language as a method for teaching reading and suggest that direct instruction in phonics can be effective.
The three studies, submitted separately but published together in the December 1991 issue of the Journal of Educational Psychology, represent the latest salvo in one of the most hotly contested disputes in education.
The reading field is sharply divided between those who advocate teaching reading by emphasizing the relationships between letters and sounds--or phonics--and those in a rapidly emerging camp who advocate the use of whole texts to stress the meaning of words.
One of the new studies--which replicated a landmark 1965 study that has become a foundation of the whole-language movement--concluded that the earlier study had overstated the benefits of reading words in context. In fact, the authors of the new paper found that good readers in their study were not better at identifying words in context than they were at reading them from a list.
The two other studies found that direct instruction in letter-sound relationships improved children's ability to identify words.
Together, the three studies bolster a strong body of research that points to the need for a balanced approach that includes both phonics instruction and the use of whole texts, according to Frank R. Vellutino, who wrote an introductory paper on the studies for the journal.
"The fact is, the research supports the utility of both code-oriented [phonics] and meaning-oriented strategies," said Mr. Vellutino, a professor of psychology at the State University of New York at Albany and the director of the child-research and study center there. "But in a system based on an alphabet, code-oriented strategies carry a little more weight."
But Kenneth S. Goodman, a professor of education at the University of Arizona and the author of the landmark 1965 study, denied that the new studies refute his analysis. Decades of research--including teachers' observations of their own students--have borne out the validity of teaching reading through the use of whole texts, Mr. Goodman said.
"Teachers aren't relying on what I said in 1965," he said. "They are relying on what their kids are doing now."
Challenging a 'Classic'
The so-called "great debate" over reading instruction, which many experts believe has escalated in recent years, has flared up in school-board rooms and legislative chambers, as well as in academic journals.
Last fall, for example, school officials in Houston permitted several elementary schools to use a drill-based phonics program after school administrators expressed dissatisfaction over a shift to whole-language methodologies. ('See Education Week, Nov. 20, 1991 .)
The three new studies provided the Journal of Educational Psychology with a "unique opportunity" to shed additional light on the controversy, said its editor, Joel R. Levin.
The first study, conducted by Tom Nicholson, a senior lecturer in the department of education at the University of Waikato in Hamilton, New Zealand, challenges the findings of one of the "classic" studies in the field.
In that 1965 experiment, Mr. Goodman asked 100 1st through 3rd graders in a Detroit-area school to read a list of words, and then asked them to read text material that included the same words that were on the lists. Examining their errors, or "miscues," Mr. Goodman found that the children made 60 percent to 80 percent fewer errors when reading words in context.
The results indicated, Mr. Goodman wrote in a response to Mr. Nicholson in a separate journal, that "it didn't make sense to introduce words in isolation before asking children to read."
Good and Poor Readers
But Mr. Nicholson, in replicating the study, notes that Mr. Goodman failed to distinguish between good and poor readers. The new study found that "the significant gains [from reading words in context] were made only by the poor and average readers and by the 6-year old good readers. The 7- and 8-year old good readers did not show significant gains."
"It seems clear that students who have strong phonics skills are better at using context than poor readers, when given the same words to guess, in easier material," Mr. Nicholson writes. "However, with more difficult material, in which words are harder to guess, phonics skills would become extremely helpful to good readers, and this is probably what enables these readers to read just as well when given words in lists as when given words in context."
Mr. Goodman, who had not read the new study, questioned its findings. He noted in an interview that previous attempts to replicate his 1965 experiment, by Mr. Nicholson and others, had erred by using "minimal contexts," rather than real texts.
In addition, he said, a substantial body of research has been conducted in miscue analysis, "and all of them support the same finding mine did."
The two other new studies tested another premise at issue in the debate over reading instruction: whether children should be taught letter sounds directly in order to develop their abilities to identify words.
Awareness of 'Phonemes'
One of the studies, conducted by Brian Byrne and Ruth Fielding Barnsley, psychologists at the University of New England in Armidale, New South Wales, Australia, examined two groups of preschoolers. One group was trained directly in identifying "phonemes," or the sounds that make up words; the other group received the training without direct instruction in phoneme identification.
The study found that the children who received the direct training were better able to identify phonemes and to recognize words than were members of the control group.
Similarly, the third study, by Barbara R. Foorman and colleagues in the department of psychology and educational psychology at the University of Houston, found that 1st graders who were given greater instruction in letter-sound identification outperformed those who were given less such instruction on a test of word recognition.
However, Mr. Vellutino pointed out that both the Byrne-Fielding Barnsicy and Foorman studies also found that many children who had received direct instruction in phoneme identification were unable to recognize many words.
These findings suggest, he said, that children must know how sounds are used in reading, as well as how sounds are represented by letters of the alphabet.
"Until you can identify words, you do not comprehend," he said. "But if you are not careful, you can make children good decoders, but poor readers."
But Mr. Goodman said such an emphasis on decoding skills reflects an outmoded view of learning. Current cognitive research, he noted, has found that children "construct" reading based on what they already know; they do not read by sounding out each word individually.
"A small group of people," Mr. Goodman said, "thinks that making sense of text depends first on identifying words."
"They talk to each other," he said. "They are not talking to teachers or to researchers in cognition."