>>This means that in relation to fault-finding on comms gear 25% of the category IV were capable of out-performing 3% of the Category I, and out-perform even higher percentages of other groups. This seems to make a mockery of the initial categorisations!<<

Not really. From my own military experience, low-IQ guys tend to grasp at solutions. They have trouble thinking methodically. They grasp and sometimes get it right. Some have enough intuition to be right more often than wrong--and those are the low-IQ guys the military likes to have which is why they have an allowance of low-IQers, but these people are rare and that is why the allowance was originally so small.

Low-IQers have to be checked on and whenever they say they completed something, you MUST check. I learned that the hard way. Low-IQers tend to be less than model sailors/soldiers also if not downright problem soldiers. Now, I did know one smart guy who was a problem and could not be trusted to do anything much less do it right but he was an exception. Most smart guys WANT to do it right and CARE that it done right even if they hate the task. Dumb guys could be good at something but, even then, their ability to advance was necessarily limited.

Moreover, the 3% of smart guys who got something wrong are far more likely to figure it out and do it right than the 25% of dopes who got it right but wouldn't have a clue of what to do if they didn't.