Your point is certainly valid--I agree we shouldn't be shoe-horning results in to fit with theory and let results speak for themselves. But I don't think there's anything wrong with proposing and performing experiments with hypotheses developed from theoretical perspectives.