Nelson A A
Am J Hosp Pharm. 1980 Feb;37(2):264-5.
The characteristics of useful research hypotheses are identified and discussed. Research hypotheses should (1) express a relationship between two or more variables, (2) have clear implications for testing, and (3) have measurable variables. Since the hypothesis determines study design, the definition of variables and the level of abstraction contained in its relationship affect the use to which study results can be put. Formulating a hypothesis is of value whenever interpretation of the results is the purpose of a study.
有用的研究假设的特征被识别和讨论。研究假设应该：（1）表达两个或更多变量之间的关系；（2）对检验有明确的意义；（3）有可测量的变量。由于假设决定研究设计，变量的定义及其关系中包含的抽象程度会影响研究结果的用途。只要研究目的是对结果进行解释，提出假设就是有价值的。
