@mhoye depends a bit on the program you're doing this for. Is this more software engineers or more mathematical computer scientists?
Anyways, probably run them through the gamut of things that software actually is used for (electrical engineers in safety-critical disciplines are sometimes a bit confused about the break things and maybe move fast philosophy of modern web-centered SE; people in modern software companies by the lack of good tooling in embedded. Aerospace by the lack of req. eng.;
@mhoye and this all leads to different company philosophies).
Then look at one or two projects that involve software development and look at what the challenges are that needed to be tackled, and how cooperation within and between teems, with vendors, libraries, users, and API consumers happens.
Look at incentive structures for different people in different constructs (FOSS enthusiast after work; PhD cand; software employee in car industry; employee at large software company; manager there),