In this prior post, I offered a possible explanation for the odd broad reading score. My discussion was specifically about four year-olds, but it happens to apply to five year-olds as well:
http://giftedissues.davidsongifted...._on_bizarre_WJ_III_resul.html#Post196822As to reading levels: they are no more reliable or unreliable on the WJIII than on most other decently robust psychometric instruments. They are just subject to the GE/AE limitations previously mentioned. In the absence of any better measure (preferably leveling/placement assessments taken from the actual curriculum to be used), the estimates are adequate as ballparks. Keep in mind that most students at a particular nominal grade level do not perform independently at the same level at which they are instructed. Also, in any given group of grade peers, there will be easily a two grade-level range, even without including significant outliers, like children with disabilities or giftedness. So a child who demonstrates achievement on a particular task equivalent to that obtained by the median member of the norm group at that grade level is not necessarily appropriate for instruction at that level.
As I've mentioned before, the GE/AE curves are also not linear, as some skills increase quite rapidly at certain grade levels, but may plateau early, or have varying rates of expected progress. Consequently, a one- or two-point difference in raw score can have a dramatic impact on GE/AE, depending on which one or two points it is.
But if all you're looking for is a general sense, then, yes, it is reasonable to assume that his instructional level in reading skills and reading comprehension of sentences and very brief passages is probably middle grades/elementary.
