So actually, there is no such thing as taking the test at the grade 10 level as a grade 5 student, unless it was scored out-of-level using grade 10 norms (which it would appear it was not; if it was, then this would become a different discussion altogether). The difference between the grades is encompassed in the design of the test basals and ceilings, with the exception of the Reading Comprehension and Oral Reading Fluency subtests, which have grade-based item sets. But administering out-of-level in the way you describe, while possibly clinically informative, would not generate standard scores for those two subtests. I suppose they could have started him at the grade 10 start points instead of grade 5, but all of those are actually called grade 9-10 start points, so it would be unexpected to call them grade 10 (and the same issue with item sets still exists).
Not sure what they considered a third of the allotted time. FWIW, it usually takes about 1.5 - 2 hours to complete. About 20 minutes of it is quite difficult to complete more quickly while also performing at a high level (i.e., usually those sections go quickly only when a student gets very few items correct; some of them are timed with minima, and others are limited by how quickly the examiner can physically proceed through the stimulus items).
To your more important question: no, there is probably no significance to achievement scores well above the scores predicted by the WISC-V FSIQ. Predicted scores take into account regression to the mean, in the form of correlation coefficients as obtained through the linking studies. As a rough example of this, say the correlation coefficients obtained for Total Achievement as predicted by FSIQ is .75 (which is around the vicinity of true). The FSIQ is +52 SS above the mean. 52x0.75 = 39. The mean (100) + 39 = 139. Hence the predicted score.
Then one should take into account the difference in scores that would be considered statistically significant and rare. A typical critical value for significance at the .05 or .01 level would be about 7 or 9, respectively, with a base rate considered to be rare (<10% or <5%) falling somewhere around 12-15 -ish standard score difference. That is, until the difference in scores (in this example) is more than about 15 points off of the predicted score, it isn't considered remarkable in either direction. That would be 154 in this case, not that far off his actual score. (And I'm not using the real tables in this case.) In any case, the point of regression to the mean is that one or the other (or both) score might have error in it that veers slightly high or slightly low, and the predicted scores try to take that into account. If both scores were more accurate than standard error would have predicted, you would get more correspondence between them than regression to the mean might have predicted.
Probably more psychometrics than you wanted to read...but the bottom line is that this isn't wildly out of line with his assessed cognitive ability. And also, he is likely up in the ceiling of the test.
ETA: should have looked back at your previous post first...you may recall that I suggested that his assessed GAI and FSIQ are likely low estiamtes of his true ability. So if his extended GAI is notably higher than the standard norms suggest, it would be even less surprising that his achievement scores are in this range. I would again suggest that (if this matters to you for access to resources or other items of value to your family) there may be utility in obtaining the extended scores from the original evaluator (or, if you do not need the numbers formally, but have the raw scores and want extended scores, pm me).
Last edited by aeh; 04/14/21 02:06 PM. Reason: thinking more
