So the phonics work at Sylvan probably helped a bit with reading, but wouldn't have as much impact on spelling, if he truly has difficulties with segmenting sounds. (And it's a little bit of a brute force method that doesn't necessarily address the underlying phonological processing issues.) It's easier to look at letter combinations and retrieve a learned sound pattern than it is to listen to a word and generate the letter patterns, partly because there are more graphemes (letter combinations that make a single sound) than there are phonemes (speech sounds). He appears to blend adequately, which means that once he learns the phonemes that go with specific graphemes, he can combine them effectively into words. On the other hand, without strong segmentation, he can't go the other direction very efficiently, since he has to be able to tease out and sequence the individual sounds before he can start applying grapheme/phoneme rules to them.
