Next in importance is choosing the desired precision of the estimates. In them depends the size of the sample we need to observe. Now we must consider the existance of factors, so called nuisance factors, that can affect the observed values. If there are any; do we need a stratified sample? or can we get along with a simple random sample?.