Example: Activity X has 80 garbage containers, of which 70 are emptied
weekly Wednesdays, and 10 are emptied twice a week on Tuesdays
Thursdays. Then monthly population is the total number of work occurrences
per month. In July 1986, there were 5 Tuesdays, Wednesdays and Thursdays.
The population for July 1986 would be 70 x 5 plus 10 x 10 or 450.
Note that the population is not 80, the total number of garbage containers.
As can be seen from the above example, when the service is scheduled, the
population for any particular month can be determined by reference to the
schedule of services. If the schedule refers to particular days of the
week, reference must be made to a calendar. The population of scheduled
services must be calculated if random sampling is used. The population must
also be calculated in connection with 100 percent inspection and planned
sampling. When services are unscheduled and random sampling is used (RSED
or RSWED), that is performed on an "as required" basis, population size must
be estimated based on historical data projected into the future.
Example: The number of service calls which occurred in a contract for
family housing maintenance at Activity Y for a period of twelve months were
Both the minimum and maximum populations need to be estimated when random
sampling is used. Population estimates may also be needed for planned
sampling and 100 percent inspection.
The minimum monthly population is estimated by finding the historical
minimum monthly population by inspection of the above monthly numbers and by
arbitrarily deducting say 20%. The minimum historical monthly population
was 570 - deducting 20% gives an estimated minimum population of 456. This
approach is conservative, but is recommended when estimating minimum monthly
populations, because failure to make a low enough estimate of the minimum
monthly population will make it impossible to validly extrapolate the
results of the random sampling.
The maximum monthly population is estimated by finding the historical
maximum monthly population by inspection of the above monthly numbers and by
adding 20%. The maximum historical monthly population was 610 -adding 20%
gives an estimated maximum population of 732. An alternative approach is to
find the average historical population which is 7101 divided by 12, which
equals 592, and add 20% giving an estimated maximum population of 711. In
most cases, the accuracy with which the minimum monthly population is
estimated will nut be critical.
The minimum population over the duration of the contract may also need to be
estimated in connection with the determination of sample sizes discussed