# Hypothesis

A hypothesis is an idea or theory which makes provisional predictions
A hypothesis states that one thing is influenced in some way by something else.
For example ""
All "things" and "something else" are variables

### Directional Hypothesis

It is extremely important to know whether an experiment has a directional hypothesis as this will affect the type of statistical test used.

One Tailed Hypothesis
When a hypothesis states a particular direction of outcome, for example "one thing increases something else" then this is called a directional hypothesis or one-tailed hypothesis
When a hypothesis states a predicted direction of outcome then it is called a directional or one-tailed hypothesis.

Two-Tailed Hypothesis
When a hypothsis does not state a direction then it is called a nondirectional hypothesis or two-taled hypothesis.
When a hypothesis doesn't state a predicted direction it is called nondirectional or two-tailed hypothesis.

Regardless of whether a hypothesis is one or two tailed it is known as a research hypothesis or alternative hypothesis.
(i.e. alternative to the null hypothsis)

### Null Hypothesis

The independent variable does not affect the dependent variable in the way we anticipated.
Also known as the hypothesis of no difference)
A null hypothesis is tested at a particular significant level.

The variable we manipulate is called the "independent variable" (IV)
The variable that will change as a consequence is called the "dependent variable" (DV) (i.e. this value depends on the value of the other variable).

Most variables can be either dependent or independent within the context of a particular experiment.

It is easier from a statistics point of view to show a difference between two sets of scores if a direction is predicted at the outset.
Rather than writing out the word "hypothesis" together with its type it is common to abbreviate this to H and add another letter or digit to signify which type of hypothesis it is:

For example H (1) or H (A) to represent the research or alternative hypothesis.
H (0) to represent the null hypothesis.