Discuss the experiment's guidelines with the other groups of observers. One of the causes of observation bias is when not all of the groups interpret the experiment in the same way. By meeting with the other groups, and explaining how the experiment should be conducted, observation bias can be avoided.
Understand the background knowledge that each group brings to the experiment. For example, one group that is conducting the experiment may be a group of students, another may be a group of researchers and another may be college professors. The students are in the process of learning the fundamentals of the material, the researchers are more accustomed to performing these types of experiments and the professors are able to connect the theories to the results.
Review the main reasons as to why the experiment is being performed. Different values can cause individuals to observe and make conclusions in different ways. For example, a person who does not want to experiment with animals may concentrate more on the animal's plight than on the experimental objective. A person who is more aware of the reason for the experiment can either choose to put her differences aside or to find someone else who can.
Make sure that each group of observers is recording data in the same way. A group who records all results is going to have better data than a group who does not. This presents both results from being presented. For example, you may provide each group with sheets where the observations can be recorded for the duration of the experiment.
Obtain detailed information about the people, animals or specimens that you are observing. Without enough information, it is easy to make generalizations that may not be true.
Do not assume that all people, animals or specimens in the experiment will possess the same characteristics or behaviors. For example, if the first 100 people in an experiment do not do well on an exam when loud music is playing, it does not mean that all people are unable to do well on an exam under such conditions.