There is a long history of discussion about the ability of researchers to generalize their findings. But findings are not the only entity that researchers can attempt to generalize. Scientists have theories, empirical hypotheses, and statistical hypotheses too. The extent to which scientists can generalize these is an open issue. As a prerequisite to proposals about generalizing theories, empirical hypotheses, and statistical hypotheses; they must be distinguished from each other. Another prerequisite is to specify, across what, the researcher wishes to generalize. I show that theories may or may not generalize across auxiliary assumptions, empirical hypotheses may or may not generalize across statistical assumptions, and statistical hypotheses may or may not generalize across inferential assumptions. The reasoning to be presented, pertaining to generalization, militates against the typical practice of using p-values to draw conclusions about hypotheses.

中文翻译：

---

泛化辅助、统计和推理假设

关于研究人员概括其发现的能力的讨论由来已久。但研究结果并不是研究人员可以尝试概括的唯一实体。科学家们也有理论、经验假设和统计假设。科学家在多大程度上可以概括这些是一个悬而未决的问题。作为关于泛化理论、经验假设和统计假设的提议的先决条件；它们必须彼此区分开来。另一个先决条件是指定研究人员希望概括的内容。我表明，理论可能会或可能不会在辅助假设中概括，经验假设可能会或可能不会在统计假设中概括，统计假设可能会或可能不会在推理假设中概括。要提出的理由，p值以得出有关假设的结论。