Will this formula work?
Can observations be viewed as
randomly sampled?
Can we estimate
s
from a
reasonably well-sampled
ICOADS period, will that be
good enough?