I'm always looking for good examples where 'obvious' control measures are not as good as people expect: could you give me a pointer/link to the relevant research providing evidence that "there is 10% chance that duplicate inspection will not catch the fault"
Here is a report just released. Take off crash of a DA42 because rudder control was reversed. Finding,
Certification of work accomplished by another person
In this occurrence, the rudder cable guide tube replacement and the installation of the rudder cables were performed by individuals other than the person certifying the work. The ACA holder who certified the work was not present during the performance of the work and was not aware of the guide tube replacement before certifying the work. Even though it is possible to inspect most of the work after it has been completed, there are some aspects of the rudder cable installation that may be difficult to validate due to their location on the aircraft, especially when the person certifying the work is not aware of the work performed on the system.
For the steps that are difficult or impossible to validate once the work is completed, it may be necessary for the person signing the maintenance release for the certification of the work, to observe the work being completed, to the extent necessary to ensure the work meets the applicable standards of airworthiness.
Finding as to a risk
If the person signing a maintenance release for work accomplished by another person does not observe the work to the extent necessary to ensure that it is performed in accordance with the applicable standards of airworthiness, there is a risk that the aircraft will not be airworthy when it is returned to service.
https://www.tsb.gc.ca/eng/rapports-r...0/a22o0060.pdf