The general rule of standardized testing is that you follow the rules. I don't think it was inappropriate of the tester to say, "The technical bulletin says to use it if there are two scores of 18 or 19, and that is not the case here." However, the language in the technical bulletin says, "are useful" not "are only used when." She probably could have gone either way.
What I think is of more concern is that a kid with Asperger's Syndrome, who might be expected to do less well on Comprehension, and who showed every sign of doing just that, had that subtest discontinued and essentially marked as "spoiled" without her actually doing so, and without that choice being clearly made and explained in the report itself.
The Wechsler manuals don't allow for, "I thought the kid was going to do badly on this one so I'm doing the other one instead." It's only if a kid really honestly spoils a subtest (e.g., I had an autistic kid once who marked "yes / no / yes / no / yes / no / yes / no" all the way down Symbol Search, and refused all of my efforts to get him to do the actual task), then you *say so* in the report. In that kiddo's case, I didn't report a score at all for that subtest -- it was a completely invalid test of his actual ability to rapidly process visual information.
But if an otherwise bright kid is just getting silly on the early Comprehension items, I would wonder if perhaps she *knew* she didn't know the answers and was trying (successfully) to distract the tester. That's different from spoiling... that's sandbagging, or at the best, plain old tanking. If she tanks a subtest because she honestly doesn't know the answers, then that *is* her score and it should be reported as such. Sorry.
I'm writing a report right now on a young adult with GT/ADHD, who has really strong achievement in mathematics but no ability to concentrate. On the WAIS (different from the WISC), for some idiotic reason, they switched from Digit-Span and Letter-Number for the Working Memory factor to Digit-Span and Arithmetic. So the kid got a rotten score on DS, rotten on LN, and ceilinged AR. I reported the WM score as calculated on the basis of DS and AR... and *then* also calculated what it would have been based on DS and LN (which included refiguring FSIQ as well), and discussed the difference and the clinical significance of that difference within the report.
The point is that you're not allowed to change the rules of the test. You're allowed to ask some "what-ifs" and to use the supplementary test information as a way to put things into perspective. That's why test interpretation is a complex process. But you can't throw out data just because it doesn't agree with your hypothesis.
(FWIW, this set of scores, including the assumed low score on Comprehension, looks quite typical of what I see in GT/AS kids on this test.)
