I wonder how to reduce the probability of a scientific manuscript to be rejected? In want context a paper should be written for it to be published on 'good' journal?
I believe the American Sociological Review has a note somewhere about what it takes to hit a top-ranked journal: Address and important debate in the field; have a good and appropriate data set to test the theories in the debate; analyze the data with proper methods; draw conclusions; sketch implications for future research. Circulate initial drafts to friends/colleagues to gather criticisms prior to journal submission and then make sure you have fixed those problems ahead of time.