When calculating the standard deviation of a dataset, should we use $\sigma$ (population SD, dividing by $n$) or $s$ (sample SD, dividing by $n-1$)? Which statements are correct? (Select all that apply)

I. Use $\sigma$ if the dataset represents the entire group of interest II. Use $s$ if the dataset is a random sample from a larger population III. The $(n-1)$ denominator provides an unbiased estimate of population variance IV. For large $n$, the difference between $\sigma$ and $s$ becomes negligible