There's a well-known story about Andrey Kolmogorov. As a student his first choice of major was not Mathematics but History, and he wrote a serious paper on taxation of Russian serfs. He notices that the amount of taxes gathered from each village was always a whole number of roubles, but each individual household paid a fractional number. So Kolmogorov concluded that the taxes were levied by the landlord on the entire village as a unit, and then the villages would divide among themselves how much each household would pay. The landlord wouldn't bother himself with fractions, therefore each village's payment was always a whole number, but the individual households' contributions weren't.
When Kolmogorov showed his calculations the advisor was not convinced about the conclusions replied: "In History we need five different proofs for every assertion." After that Kolmogorov changed his major to Mathematics "because in Mathematics you only need one proof" and went on to develop modern Probability Theory and many other contributions to the subject.
Therefore the question: How many proofs are required for an assertion about a historic fact or trend would become generally accepted? What are the modern criteria for veracity of a historic fact?