It means that you design things so that the variable cannot influence experimental results. It can be inherent to the experiment, or it can be part of your analysis.
So if I control for gender during analysis, I might, for instance, only compare participants of the same gender to one another when doing statistical tests, etc.
An example of control inherent to the experiment would be testing a drug only on patients who have had a certain type of cancer for exactly 3 months - you would be controlling for diagnosis.
No comments:
Post a Comment