My advice is to pick a topic that a) you're interested in and B) is feasbile, then do some research about what's already been done in that area and think about how you could contribute to the existing literature. Remember to take ethics into account and also that you won't get funding and will most likely only be able to use the student population for data collection.