I agree more with Post 2 because I think it is more in line with what you are probably talking about (rather than a scientific experiment as in Post 3).
Your hypothesis might also be something like "One of the biggest problems for soldiers on a day to day basis is boredom." That would give you something to try to prove in your discussion of what soldiers do in a typical day.
For a hypothesis, you have to take something that has to be tested. You don't need to know what the conclusion would be, nor should the attempt to prove the hypothesis right or wrong create a bias while you are testing the hypothesis.
As an example, your hypothesis could be something like "Exposure to excessive violence causes soldiers to sleep less than people with other professions."
You can test this by considering people in other professions who also lead a life with a lot of exposure to violence, policemen for example. If you find it true that soldiers do sleep less than others, you would have to investigate if it is exposure to violence that is causing this or there is some other reason. Perhaps they are trained to sleep less, or their requirement of sleep has been reduced due to being away from their home and family, etc.
the problem could be condtions, weather, being away from home, chance of death, seeing death, losing freinds.
Hypothesis could be somthing like, that the average daily life of soldiers can cause deep emotional impacts even after there return home