The wild scatter is definitely something I'd worry about. However, I wouldn't be too sure as to whether the scatter is valid. Given the Flynn effect and that the WISC-III is almost two versions outdated since they're moving onto the WISC-V in the next year, I believe, I wouldn't personally be too confident that the high scores aren't a bit inflated as well making the scatter perhaps lesser than it appears at face value.

I'm not familiar with the WJ-III cognitive, but am wondering if you've looked @ the SB-V b/c it isn't timed and might, thus, be better for a child with a processing speed deficit.

Study Strategies for Accelerated Learners