After just having spent 4 years in the US where most, if not all science is very hypothesis-driven and people tend to work on a specific question it is interesting to see how different that is in some labs here in the homecountry. The other day I witnessed a conversation between a scientist and hir department head. The scientist works on individual differences in bunny hopping (to steal this analogy from DrugMonkey) and was planning to write a proposal studying individual hopping preferences in relation to food collection since that follows nicely from what this scientist had been working on for years. However, since this scientist is not a department head, ze cannot submit this proposal hirself (yes, let's not even go there to discuss this…) so the dept head has to submit. And the department head thought that perhaps what reviewers are interested in now is not bunny hopping but giraffe running. And not in relation to food collection but in relation to the giraffe's circadian rhythm. The scientist was unhappy, as ze wanted to pursue the questions that ze had been working on for years. But since the dept head was the one submitting the proposal there was really not that much ze could do here.
And you might wonder: wouldn't the scientist and the dept head need preliminary data on giraffe walking and circadian rhythms? Apparently not. I have seen people get grants working on something they have NEVER done before just because their question apparently sounds interesting and the reviewers for some reason have faith they will be able to do it.
So when this scientist gets funded, ze may need weeks, months or even years to set up a paradigm to study giraffe walking and characterize its circadian rhythm. Or frantically look for a postdoc who is experienced in doing this. But apparently that's fine by the funding organization.