抄録
In this paper, the rank test using special score [=2i-1(i=rank)] is described.
This test has the advantage in that we can calculate the value of probability of significance exactly, even though the size of sample is large, and we can select the level of significance more freely than any other test.
The power of this test is examined by computer, and is compared with the power of Wilcoxon test, Siegel-Tukey test, and Savage test. Consequently it is proved that this test is effective for location and scale tests.
