Saturday, October 8, 2011

Learning Words as Whole: An Experiment

In Reading in the Brain: The Science and Evolution of a Human Invention, Stanislaus Dehaene, describes an experiment conducted by Bruce McCandliss at the Sackler Institute. McCandliss created a new alphabet and used it to write words. But to create the words he stacked the letters one atop the other, rather than placing them side-to-side. He told two groups of students the meaning of 30 of the words. He then asked one group of students to "memorize each word as a global shape," without informing them of the existence of letters. He informed the second group that the words "were made up of a sequence of letters written from bottom to top."
Thereafter, the two groups were trained identically through repeated exposure to a given shape and the corresponding English name.
The very slight difference in instruction given to the two groups had an impressive cascading impact. After one day of training on a list of thirty words, the whole-language group was actually better at recognizing them than the analytic group, who were still struggling to discover the letters. This interesting result fits with reports by many advocates of the whole-language method. They repeatedly assert that their approach gives children a head start. However, this is only true at the beginning, for the first thirty words or so. The acquisition of letter-to-sound correspondence requires greater initial effort, but the results in the long run are more rewarding. Indeed, on the second day, when students learned a new list of thirty words, the whole-language group began to lose ground. They learned most of the new words - but at the expense of the initial list, which they quickly forgot. The same pattern recurred each time a new list was introduced....The group was trying to accomplish the impossible task of learning each word independently. (pp. 225-226)
Dehaene provides further explanation of the experiment and the results.

Although there are exceptions, many, many teachers, paraprofessionals, and volunteers are currently drilling students in classrooms across the country with lists of whole words, without initially working with the students to analyze the letter-sound correspondences therein. In kindergarten classrooms this routine is typically focused on the 25 or so "sight words" that the students "need to learn" by the end of the year. This is done in the name of "fluency." [I have yet to understand why this type of drilling is not included in the "drill and kill" accusation thrown at phonics proponents.] The students who have not mastered the letter-sound correspondences continue to confuse these high-frequency words with each other... so the drills continue. "They need to recognize these words as wholes," is a common statement. It is terribly unfortunate, for the teachers and for the students, that the scientific research into HOW such words most easily become recognizable by readers has not trickled down to the classroom.

The book, now out in paperback, is available at local independent bookstores and at:
Powell's Books and Amazon

No comments: