The pickup of garbage from 60 containers at various locations on the station
was monitored on each pickup day for a period of one month. The garbage was
required to be picked up before noon on the scheduled pickup days. The
results of the inspection were as follows:
A. Response time
The defect rate (DR), when planned sampling is used, is calculated by
dividing the total number of observed defects (including those detected by
validated customer complaints and incidental inspections) by the total
population of services so that
DR = Total Number of Observed Defects x 100
Dividing the number of defects in the sample by the sample size gives a
defect rate for the sample only, and the use of this "Observed Defect Rate",
(ODR) to either assess the contractor's performance or take deductions to
the contractor's price is contrary to NAVFACENGCOM policy. This is because
neither the size of the sample nor the method of sample selection are in
accordance with the rules for statistical sampling techniques. As a result
the extrapolation of planned sampling has no validity and the ODR may not be
compared with the MADR, the MADR being defined as the defect rate in the
population above which the contractor's quality control is considered
In this particular example, the DR (0.6%) exceeds the MADR (0.2%) so the
contractor's quality control can be rated as unsatisfactory.
Changes in the DR from month to month give a reliable indication of whether
the contractor's quality control is getting better or worse provided the
same sampling procedures are followed each month.
It is not always possible to rate the contractor's quality control using
planned sampling, especially when small sample sizes and relatively high
MADRs in the 5% to 10% range are used.
An alternate approach is to specify the MADR as a discrete number of defects
rather than as a percentage when planned sampling is the primary method of
surveillance (e.g. 15 defects per month rather than 0.2%).
(c) Validated Customer Complaints. When validated customer complaints
is used as the primary sampling method, it is generally preferable to