Estimate the range likely to contain the next single observation.
x̄ ± t·s·√(1 + 1/n), with df = n − 1.