is inversely related to random error. • Small sample sizes and few observations generally widen the confidence interval around an estimate of an effect, decreasing the precision of that estimate and lowering any rating of the quality of the evidence.