Use of confidence intervals to assess precision of normal range estimates

1. When data are obtained for estimating a normal range, their efficacy may be assessed by deriving confidence intervals for the normal range endpoints (3,4). If the intervals are narrow, the normal range has been estimated precisely. If they are wide the data are not sufficient to precisely estimate the normal limits. 2. In this report confidence interval calculations are applied to data from a recent study of norms by sex and age for the SMA 12/60 screening group of tests (5). Some explanations are given for those instances where the derived limits are wide.