We don't know the size of the sample for the preliminary estimate, either.
With a little thought, perhaps we can get past it.
This estimate, 1,497 seems like a HUGE number. It may be, but in the absence of additional information, that may be the choice.
1,497/0.05 = 30,000 or so. If the university is big enough, 1,497 is not so bad. It's not like we're conducting destructive experiments or asking lots of questions. 1,500 surveys on the street is not so bad. It could take a while, and I'm not sure how you would control duplicates, but why not.