ABSTRACT
There can be significant pressure from all kinds of directions to pinpoint a cause. Early investigation can typically show that a system behaved as designed or programmed, and that there was nothing mechanically wrong with it. This is taken as automatic evidence that the mishap must have been caused by human error, after all, nothing was wrong with the system. Human error causes failure in otherwise safe, well-functioning systems. Some investigators get so frustrated with the need to present a 'probable cause' that they simply pick the last technical event that, if removed, could have prevented the mishap from occurring. They end up with ridiculously short 'causes' that are almost impossible to refute because that particular factor was certainly necessary to have the outcome occur. Sufficiency is no longer a criterion for including a cause or contributory factor, as it would generate a causal list at least as long as the investigation report itself.
