Asymptotic relative efficiency.

Asymptotic Relative Efficiency ] was pleased to read "Promise in Nonparametrics" by Royeen and Seaver (March, ] 986). As the authors state, analysis of data by non parametric methods when the strong assumptions for parametric counterparts cannot be met provides the researcher with effective statistical procedures that may be more realistic and valid in a given situation. It is also interesting to note that a measure of relative efficiency, called the asymptotic relative efficiency (ARE), can provide for a comparison of a nonparametric test with its parametric counterpart under normal distribution assumptions (1). For example, the ARE f,::>r the nonparametric Spearman rank correlation (when compared with its parametric counterpart of the Pearson product-mc-ment correia tion) is .912. This is interpreted to mean that with N = 100 using the Spearman, you will obtain the same efficiency as if the Pearson had been used with N = 91. This index may possibly assist in the choice of a nonparametric technique if more than one is appropriate in a given situation. The interested reader is referred to Gibbons for measures of asymptotic relative efficiency of other nonparametric tests. Susan Heling Kaplan, MHS, OTR Miami, Florida