Use repeated trials to judge scatter

Use repeated measurements to separate a best estimate from scatter in the data. Calculate or reason from the mean and spread, then describe why repeated trials reduce random effects but do not automatically fix every problem.