This paper presents the information on the characteristics of a good research hypothesis. It explores and explains the differences between a non-directional research and a directional research hypothesis, and observes what distinguishes a null hypothesis from a research hypothesis.

The major characteristics of a good research hypothesis include as follows:

- A good hypothesis should rely on sound reasoning;
- A good hypothesis neatly affirms the relationship between variables;
- A good hypothesis offers a valid explanation for some outcome;
- A good hypothesis should be under test in a valid period of time;
- A good hypothesis should have an if–then statement.

A null hypothesis is a basic testable assumption, which is usually stated as the absence of differences, the lack of effect, etc. In turn, Delampady emphasized that “one of the major justifications for testing a point null hypothesis is that it can be considered as an approximation to an appropriate interval hypothesis in a large number of situations” (120). A research hypothesis is a methodological characteristic of some research, a scientific hypothesis advanced to explain any phenomenon that requires some verification by experience to be a reliable scientific knowledge. A hypothesis differs from a null hypothesis as well as simple assumptions in a number of features. These include:

- Verifiability;
- Applicability to the widest possible range of phenomena;
- Relative simplicity.

The major differences between a non-directional research hypothesis and a directional research hypothesis are that a directional hypothesis is considered to be one tailed. It is possible to state that through manipulating an independent variable there must be some change in a dependent variable.Thus, it is possible to guess if that change will be negative or maybe positive. In turn, a non-directional hypothesis can be described as two tailed. With the help of manipulating an independent variable there must be a reasonable change in a dependent variable, but it is impossible to guess if that change will be negative or positive.

In conclusion, it is possible to say that a hypothesis is an unproved assertion or assumption. Any hypothesis must be refutable. Irrefutable assumptions (for example, axioms) cannot be considered to be hypotheses.

Works cited:

Delampady, Mohan. “Lower Bounds on Bayes Factors for Interval Null Hypotheses.”Journal of the American Statistical Association 84.405 (1989): 120-124.