The precision of an effect estimate is a quantification of the uncertainty due to the play of chance around the effect estimate.
It is usually expressed as a 95% confidence interval. An effect estimate can be precise (i.e. have a narrow confidence interval) but biased (i.e. distorted away from the actual effect because of systematic errors).
Conversely, an effect estimate can be unbiased, but imprecise.
If you feel that this definition hasn't helped you to understand the term, click on our monkey to let us know.