Wow.  That is a difficult one, but you could try behavioral observations (e.g., drinking, smoking, eating) all things you could observe.  Lots of questionnaires assess these things, but they are also commonly measured it the lab.  You could also try something like behavioral measures of concentration or problem solving or something to try to index distraction-type coping.  In any scenario, you'd have to be very careful to clearly isolate what is "coping" from what is stress outcomes.  All these things could potentially be considered either.  Your design would need to be strong and clear: stress-->coping (eating for example)-->bmi or weight gain.

Hope that helps.

 Thanks Loren, the idea of behavioral measurement is fantastic.

Thanks Chris, we are going to measure GSR, heart rate response and EEG. But we cannot offer any other sensor technique. regards.

I think design issues are paramount for you.  Be sure you can clearly tell what are stress outcomes and what are coping responses.

Thanks Loren, this is helpful.

I would suggest you look at pivotal research done on this topic in the early 70-s and to the present by many researchers doing work for various military organizations.  It is still relevant and often overlooked, leading to repeating what we already have learned.  Just a suggestion ...

Thanks Barbara, interesting idea.