Usually, two statistical procedures A and B are compared by means of their asymptotic relative efficiency e. If e= 1, however, it is more informative to compare A and B by means of the concept of deficiency, which was introduced by Hodges and Lehmann . In the present paper we use this concept for the comparison of linear rank tests and parametric tests for the symmetry problem. In this problem, the hypothesis has to be tested that a sample comes from a distribution that is symmetric about zero. The results provide new and strong edivence for the nice performance of linear rank tests for the symmetry problem. The present paper gives a survey of the results obtained by Albers, Bickel and van Zwet  and by Albers .
|Publication status||Published - 1975|