I would start from the following checks for human error, some of which are very basic, but you'd be surprised how often even good practitioners have a momentary senior moment...
1. age/birthdate and date of administration
2. form (A or B) on both the paper record form and in the scoring software. And make sure all the subtests, record forms, and test easels were from the same form.
3. norm (original WJIII or WJIIINU 2005 normative update)
4. age vs. grade norms (in OP's case, these should have been age norms, as there are no preschool norms)
5. stray spaces or x's in computer entry
6. entered in correct fields, or with correct item set labels
7. reading fluency score entered correctly (it should be correct items minus incorrect items, no penalty for omissions)
These are the top errors from my personal experience, and from the scoring FAQ on the Riverside website.
Just as a general comment about the WJ family, the composite scores (Broad ---, Brief ---, etc.) are not simple averages, or even close. Different subtests are weighted differently. Likewise, the clusters are weighted differently to develop the GIA (on the COG). There will occasionally be weird results that are legit (which is why I am not as concerned about the Brief Reading).
Last edited by aeh; 06/26/14 02:40 PM. Reason: that would be correctly, not incorrectly!
