## 2. Overview of hypotheses

How does the process of hypothesis generation work? A 'research hypothesis' is a plausible explanation for a phenomenon or observation. It should logically provide one or more predictions, which can then be tested. Testing of predictions could involve making further observation and measurement, or could assess whether the predictions are compatible with other established facts related to the current observation.

A research hypothesis is a deductive approach to a problem. It generates predictions about the problem, which can be tested. In the formal testing of predictions using statistical methods, a 'statistical hypothesis' is defined that provides a rigorous test of a prediction. In Section 4, you will see that the statistical hypothesis is typically a 'null hypothesis', that is the prediction is false.