MartPigFace:
"...just need to check who's able to think!"
In our lab at school, there was a sign on the wall that read, "One valid test is worth a thousand expert opinions."
Maybe subconsciously this thought in part led me to my first post on this thread (20 Mar 05 19:15). In light of your specific needs, the interviewee's answers would allow you to weed out those who do not understand this critical principle. You could taylor your questions so that the bright one(s) will ask appropriate questions about the component/product mfg process, application, and field survival/failure history before they respond with answers for what they consider valid testing.
One of my tenets is that an accurate definition of the problem is a large step in determinig the solution. For example; experience in failure analysis helps me to design for reliability. Show the applicant some failed parts that exihibit classical failure modes and have them identify the failure mechanism. you could lay several failed parts on a table/desk and ask them to choose which one exhibits fatigue, single cycle overload, wear induced failure, etc. Does this pattern exhibit torsional, bending, combined stress? You could select some excerpts from failure reports including photomicrographs, case hardness microhardness traverse, material qualification test reports, etc and ask them to explain the meaning of the results, microstructural anomalies, etc.