Roger,
I wasnt suggesting any Homer Simpsonisms for yourself, it was a poor attempt at incisive wit on a bored afternoon.
The serious point is, if non-technical skills are as critical to the safe outcome of a sortie as are technical skills (a fact borne out by LOSA data), then we should be looking at them during selection, line training, recurrent training, and indeed, all stages of the training process. Of course, you cant prove a negative, so if someone does not throw a wobbly on one occasion in the simulator, you cannot de facto say that they are necessarily an all-round good egg. But, if we maintain good training records, hopefully a pattern will emerge if someone has a tendency, say, to rush their checks and miss items, just as if they have a tendency always to be too high on the approach. In this way we can weed out the wrong candidates during the training process IF they display poor CRM skills, just as if they displayed poor handling skills. Of course you cant artificially test all aspects of someone's character, but if you put a crew under a bit of pressure, you can see if they are capable of working together effectively as a team, whether or not they like each other, are offended, etc etc. You can see their behaviour, and make an objective assessment about it.