If your sample size is 1500 from 2000 total then the math is likely to be good. If it's from 4.5 billion (for example examining the rate of world-wide diseases but only examining the Chinese) then it may be suspect. What are you trying to establish with your sample and what is the size of the overall population?