摘要：Presumptive Statistics that clarifies characteristics of a population using those of random samples of unlimited population is necessary for our researches. At present, presumption and certification representing major methods of this statistics are easy to be conducted by virtue of development of softwares for the statistics. On the other hand, in spite of this handiness, they also involve a risk of improper use without understanding the meaning of the method. Especially, although certification of difference in average values and analyses of correlation are frequently observed in articles, it is seldom that the results obtained were sufficiently discussed. This tendency is considered to result from lack of understanding of the concept of presumption or certification, not from that of meaning. Comprehension of the concept of presumptive statistics is indispensable to consider what kind of or to what accuracy presumption or certification should be performed at the preparation stage of the experiment preceding sampling.
