Isn't that part of the scientific process? Is there some other way to come to the truth aside from observation and examining variables related to the observed outcomes?