Randomization is a critical aspect of any experiment. In almost all cases, the population being studied is much too large to study every individual, so a sample of the population is studied with the assumption that trends and relationships seen in the sample are also present in the population as a whole. In order for this to be a good assumption, the sample must be completely random in order to eliminate any bias towards a specific type of individual.
In an ideal world, all samples would be completely random, but this is not logistically possible in many cases. For example, at Staffanson Prairie Preserve, there are thousands of Echinacea that bloom every year. It would be a near impossibility to visit every single plant or even to select a completely random sample of plants within the preserve. For this reason, the Echinacea Project created a 10-meter wide transect through the preserve and studies the plants that fall within this transect. While this is not a truly random sample, it is able to approximate the range of conditions seen throughout the preserve.
Another example of randomization is something I’ve been working on in the lab for the last week. Many of the heads contain several hundred achenes, so x-raying all of them to determine whether or not they contain a seed would be extremely time consuming and difficult. In order to simplify the process, I am randomly selecting 1/6th of the achenes from each head in order to estimate seed set for each head. While this will not give me the exact seed set, it will give me a very good approximation that will be sufficient for our analyses. Pictures of the randomization process are shown below—achenes are randomly dispersed on a wheel divided into twelve labeled sections of equal size. Next, two letters are selected from a list of random letters, and the achenes that fall within these sections are selected to be x-rayed.
Achenes on the randomization wheel
Randomized achenes–labeled and ready for x-raying
