I have gone through tonnes of online resources (other stackexchange postings as well) and textbooks but still have not been able to grasp the concept. Most of them have vague explanations like "if you want a mean fixed, only n-1 observations can vary and not n". Does anyone have a better explanation to it? For standard deviation alone, the n-1 is also explained as a Bessel's correction which does make sense. But what about degrees of freedom in general or in other statistics?

| cite | |