If we believe that a certain thing ought to be true, we can almost always find either an instance where it is true, or someone who believes it ought to be true. It is ever so hard when a concrete fact illustrates a hope to weigh that fact properly. When the first six people we meet agree with us, it is not easy to remember that they may all have read the same newspaper at breakfast. And yet we cannot send out a questionnaire to 816 random samples every time we wish to estimate a probability. In dealing with any large mass of facts, the presumption is against our having picked true samples, if
...more
