Krashen on phonics and read-alouds

Posted on EdWeek Website, following Early-Literacy Findings Unveiled

http://www.edweek.org/ew/articles/2009/01/08/18read.h28.html#comments

The NELP (National Early Literacy Panel) study appears to have come down in favor of phonics and was apathetic about read-alouds.
This sounds familiar: The National Reading Panel, we were told, came down in favor of phonics and was apathetic about self-selected reading in school. A close look at the National Reading Panel by Elaine Garan in a Kappan article in 2001 questioned the phonics conclusion, and in another Kappan article, also in 2001, I questioned the conclusion about self-selected reading in school.
A close look at NELP gives us reason to question their conclusions about phonics and read-alouds.
The claim: Code-related interventions are the most effective. Previous research has shown that heavy code-emphasis approaches are indeed effective on tests of decoding, that is, tests of pronouncing words outloud. NELP found this also.
Previous research has also shown that heavy code-emphasis approaches have very little or no impact on tests of reading comprehension given after grade one, tests in which students have to understand what they read. Reading comprehension tests given to very young children, as the National Reading Panel has noted, "generally use extremely short (usually one sentence) 'passages.' On these short passages, the effects of decoding should be strong" (National Reading Panel, 2000, p. 2-115). (For discussion, see Elaine Garan, 2002, Resisting Reading Mandates, page 15).
The only reading comprehension tests included by NELP were given in grade one or earlier.
NELP thus does not address the important question of whether heavy decoding approaches really help children learn to read, if we define reading as understanding texts.
The claim: Reading aloud to children is not as effective as previously thought, having only a moderate effect of oral-language and knowledge of print features.
Read-alouds (actually "shared reading interventions") showed a fairly strong effect on oral language in the NELP analysis, with most of the impact on vocabulary (d = .6). Shared reading interventions did not relate to phonemic awareness or knowledge of the alphabet, but that is not how read-alouds are claimed to help reading.
Jim Trelease (2006, The Read-Aloud Handbook, sixth edition) states the case for read-alouds succinctly. Reading aloud:
1. conditions the child's brain to associate reading with pleasure
2. creates background knowledge
3. builds vocabulary
4. provides a reading role model.
NELP confirmed number (3), and did not consider the others. In other words, read-alouds help familiarize children with the special language of writing, and gets children excited about books, inviting them into the literacy club (Frank Smith, 1986, Joining the Literacy Club).