Note: a question similar to that I'm going to ask was discussed here, but answers to the questions I'm going to ask were not covered there.
Basically, I'd like to understand how significant is the gap between lower- and higher-ranked graduate schools in mathematics. I'll be referring to the USN graduate schools ranking. Let me fix the notation straight away. By 'higher-ranked schools' (which I will also refer to as 'top schools') I will mean top 10 (or if you want top 20) and by 'lower-ranked schools' I will mean schools ranked 20-40 (all according to USN).
What I understood from the question I provided the link to above is that being amidst 'top students' at a higher ranked school is more beneficial than being amidst 'average students' at a lower-ranked school. Also, higher-ranked schools may be more diverse and provide more opportunities in meeting people from other top schools as well as in obtaining a job in the academia.
But I still have a couple of questions.
The first one is about the first two years of the study. Is, in general, the instruction level at lower-ranked schools worse than that in higher-ranked ones? Also, are graduate corses at 'top schools' more difficult to master and to pass? If so, does it imply that one needs a better preparation (i.e., a stronger mathematical background) to succeed in a top graduate program? Also, does it imply that the students enrolled in a top program will eventually have a better mathematical background that will make it easier for them to conduct research? If you have anything else to say about the coursework at top universities in comparison to that at lower-ranked universities, I would appreciate it.
Secondly, in the question I referred to above (or elsewhere), some people mentioned that exposure to new ideas in various branches of mathematics (which is one of the advantages of top programs) is a consequence of the size of the department and the 'quality' of faculty/post-docs/students. Whereas I do agree that students at top universities are more knowledgable and creative, I cannot see why the other assertions hold. Correct me if I am wrong but the the math department of say Stony Brook or Indiana (ranked 25 and 34, resp.) is not smaller than that of Chicago or Columbia (ranked 5 and 9, resp.). Furthermore, the vast majority of professors in all of the places mentioned are alumni of Harvard/Princeton/Berkeley/MIT/Chicago/Stanford (i.e., a top school in my terminology); post-docs also come from very prestigious places to all of the four mentioned universities. So what makes Chicago or Columbia 'better' than Stony Brook or Indiana? Just their name?