Both methods give the same variance; the choice of divisor does not matter for any dataset.

Dividing by n gives a population variance of 31.17; dividing by n−1 gives a sample variance of 37.4, which is an unbiased estimator.

Dividing by n−1 always gives a smaller variance than dividing by n.

The choice between n and n−1 affects only large datasets (n > 50), not small samples.