I'm not sure what you mean by 'statistically' above. In general you can't say whether a sample size of 2000 will get you within a specific margin of error with

I'm not sure what you mean by 'statistically' above. In general you can't say whether a sample size of 2000 will get you within a specific margin of error without additional information. It very much depends e.g. on what the hypothesis you're testing is or what you're trying to estimate.