Small Sample Properties of the Spearman Autocorrelation Estimator

Abstract
The small sample properties of a variant of the Spearman rank correlation coefficient applied in the time-series context were investigated through Monte Carlo methods. The rank method ( r1S) has even greater bias than the highly biased conventional parametric procedure; a traditional test of H0: ρ1 = 0 based on ( r1S) yields unacceptable properties. Empirical small sample distributions associated with the rank coefficient differ markedly from the distributions predicted by asymptotic theory. It is concluded that neither rank nor conventional parametric estimators and hypothesis tests are appropriate for very small samples in applications of time-series analysis that have been recommended in the behavioral and social science literature.

This publication has 4 references indexed in Scilit: