Experimental Design

Eliminating Bias

There should be no systematic difference between treatment groups. Bias could lead to false positive or false negative results. Examples of where bias might occur include:

• Treatment groups being kept in different environments.
• Measurements on each treatment group taken at different times or by different people.
• The unintentional favouring of one group when the experiment is not done blind.
• Failure to use randomisation throughout an experiment.

Blind: An experiment is done "blind" when the experimental units are coded so that the people conducting the experiment do not know to which group an animal (or other experimental subject) has been assigned. It is a way of avoiding bias.