They use a behavioural style of interview. For fresh graduates coming from the job fair, they pack you all in a hotel, near one of their production mills and in groups, give you a challenge to work out as a team, and observe how you work in the group. The whole process used to work like the NFL Draft (in fact, it was called the Draft - where after day one you were given a sheet and you were asked to rank order your locations of preference to work at within the organisation, and the interviewers/reps from the various company locations would then negotiate amongst themselves about offering you a position based on your 3 location interviews and, pretty much, how much that location wanted you. Having studied Game Theory in grad school, I understood why R&D exited this Draft process, but as I understand still continues amongst the mills. Why do I say that, at least as it relates to R&D? Well, first, those who go into research have a, shall I say, eccentric personality that may not lend itself well to selecting amongst what, amongst the manufacturing centres, amounts to a contest of picking the best alpha. Second, having studied Game Theory, this type of selection process is inefficient as it leads to misallocation resulting in efficiency losses: specifically when an applicant ranks their location while strategizing where to go based on where they perceive they will be hired and not where they prefer to be, this does not lead to a mechanism that is "truth telling", that is, one in which actual preferences are revealed. In short, this mechanism leads to misallocation of labor based on the applicant's perception of where they would be best hired and not where they would be best placed.