Whenever psychologists carry out a study, they must start with a hypothesis: a prediction of what they will find. This is usually an intelligent guess based on an idea or theory which the experiment is designed to test.

The experimental hypothesis is sometimes referred to as the alternative hypothesis. Remember that some investigations are not experiments (they may be observations, correlations, etc.) and in this case we do not start out with an experimental hypothesis but with an alternative hypothesis. When we carry out an experiment we can use either term.

Let us say that I believe that eating cheese at bedtime affects the number of nightmares you will have.

Condition 1 - participants eat cheese before they go to bed.
Condition 2 - participants eat no cheese before they go to bed

The experimental hypothesis can either predict that a difference between conditions in one direction (directional hypothesis) or in either direction (non-directional hypothesis).

Directional Hypothesis (1) - people who eat cheese before they go to bed will have significantly more nightmares than people who do not eat cheese before they go to bed.
or
Directional Hypothesis (2) - people who eat cheese before they go to bed will have significantly less nightmares than people who do not eat cheese before they go to bed.

Non-Directional Hypothesis - there will be a significant difference in the number of nightmares experienced between the two groups (those who do and do not eat cheese before they go to bed).