Section Three--Cautions Regarding the Teaching of Phonemic Awareness and Phonics
The National Reading Panel did not find that phonemic awareness and/or phonics must be taught first, before children begin to read and write.
There is, in fact, no evidencein the NRP Report or elsewherethat children must develop phonemic awareness or phonics before they begin to read print. The panelists simply found that phonemic awareness and phonics training had the most effect if children had not already developed that knowledge through reading (Subgroups, p.2-33, 2-43, and 2-133).
The National Reading Panel did not find that phonemic awareness and/or phonics should be taught in isolation.
Repeatedly, the point is made that phonemic awareness is best learned when taught in conjunction with letters (Subgroups, p. 2-33 and 2-34). Furthermore, phonemic awareness training is more effective when children are taught how to apply PA skills to reading and writing tasks (Subgroups, p. 2-40). With regard to phonics, the NRP did not find evidence to conclude that teaching phonics in isolation is better than teaching it systematically in context (Subgroups, p. 2-132).
The NRP did not find that the benefits for teaching phonemic awareness and phonics are lasting.
With regard to teaching phonemic awareness, there was no long-term evidence from the studies considered. (Also, Krashen  finds no long-term effects.) With regard to teaching phonics systematically, the slight advantage for scores on comprehension disappeared after first grade, when the comprehension passages on standardized tests became longer than one sentence (Subgroups, phonics, Appendix G, pp. 2-169 through 2-176). Furthermore, teaching phonemic awareness does not ensure that children will learn to read and write (Subgroups, p. 2-43).
The NRP found that a small amount of phonemic awareness training is better than a lot.
The NRP found that transfer to reading was most successful when a program of phonemic awareness was taught for fewer than 20 hours, with sessions lasting no more than 30 minutes (Subgroups, p. 2-42). In fact, effect sizes were more than twice as large for shorter programs than for the longest-lasting programs (Subgroups, p. 2-42). Furthermore, teaching that focused on only one or two ways of manipulating sounds was more effective than teaching that focused on more (Subgroups, p. 2-28). Regarding phonics, the NRP made no determination about how much teaching of phonics was maximally effective.