Similarly, if we had one or two height observations that were 60 cm from the mean, we might disregard them in certain
analyses as resulting from a process different from the process producing the rest of the data (e.g., from a mutation or birth defect).