Precision

Precision is the flip side of the width of a confidence interval: the narrower the interval, the more precisely you have located the parameter. You buy precision with sample size, and the price rises quickly, because the margin of error shrinks like $1/\sqrt{n}$. For example, going from n = 100 to n = 400 doubles $\sqrt{n}$ from 10 to 20 and so halves the margin of error, turning a 6-point interval into a 3-point one. Precision is not accuracy: a biased sampling method can hand you a very precise estimate of the wrong number.