Method
Various methods have been detailed in the relevant study guides.
A method is planned so the prediction can be tested.
It is sometimes appropriate to setup a control experiment.
A control experiment is necessary for comparison purposes and proves that any change in the results is due to the independent variable being altered.
A method needs to produce accurate, reliable results.
The accuracy of a measurement is dependent on the quality of the measuring apparatus and the skill of the scientist involved.
For data to be considered reliable, any variation in values must be small.
Repeating a scientific investigation makes it more reliable.
Example
Practical B2 - Investigating the energy content of food
Accuracy can be improved by using a balance to measure the mass of food.
Reliability can be improved by completing the experiment for each food more than once and calculating an average temperature increase.