evaluation criteria. Evaluation procedures should include provisions for
recording the successful accomplishment of rework by the contractor.
(j) Analysis of Results. Analysis for random sampling consists of
calculating the Observed Defect Rate (0CR) for the surveillance period and
then calculating the Defect Rate (DR) for the entire population of
services. The ODR is representative of the defect rate in the sample and is
calculated as follows:
Number of Defects
Number Evaluations Conducted (Sample Size)
The ODR is extrapolated to determine the defect rate (DR) in the population
by deducting the adjustment factor (AF) given in Appendix F, Table 3 from
DR = ODR - AF = Number of Defects in Population
Defects outside the sample which may have been found through validated
customer complaints or incidental inspections cannot be included in the
number of defects for the purposes of this calculation.
4-750 Validated Customer Complaints. A validated customer complaint is
any customer complaint identifying a contract or defect that the QAE has
validated by firsthand information from the site. Complaints not validated
may not be used.
(a) Contract Requirement. List the contract requirements together with
w o r k r e q u i r e m e n t s o f p e r f o r m from the expanded performance
requirements summary (PRS), see Figure 4-1.
(b) Primary Method of Surveillance. The primary method of surveillance
is validated customer complaints. Incidental inspections may be used as a
secondary evaluation method.
(c) MADR. The Maximum Allowable Defect Rate (MADR) is specified in the
PRS. The MADR is the defect rate above which the contractor's quality
control is considered unsatisfactory. When the use of validated-
complaints is the prime surveillance method, MADRs should be stated in terms
of defects per surveillance period.
(d) Identification of the population. Define a single occurrence of
work and determine the number of occurrences to be performed during the
surveillance period. Sequentially number the work locations in accordance
(e) Sequential Numbering of the population. This is generally not
(f) Level of Surveillance. This is not applicable.
(g) Determination of the Sample Size. This is not applicable.