The Risk Baseline Analysis Screen, which is activated by
selecting Baseline on the Main
Screen menu, can be used to define maximum false accept and false reject risks for a given application, corporate policy requirement, contract requirement or quality objective.
Accuracy Ratio Data
Sometimes, as with ANSI/NCSL Z540 and ISO 10012, a measurement decision risk baseline is expressed as an accuracy ratio requirement.
In the case of Z540 and ISO 10012, the requirement is established primarily to control false accept risk. The wording differs between the two documents, but the sentiment is the same. The risk baseline is stated
as: "The total uncertainty in a measurement process shall not exceed 25% of the tolerance limits of the parameter being measured."
The term "total uncertainty" is usually taken to include both measurement process uncertainty and measuring parameter bias uncertainty.
Baselines may also be express directly in terms of risk. For instance, if an organization is committed to holding false accept risks to less than one percent and false reject risks to less than two percent, this commitment comprises a measurement decision risk baseline.
False accept baselines are set to limit the risk of passing "unacceptable" parameters. Obviously, the level of acceptable false accept risk varies from situation to situation.
The acceptable level of false accept risk also varies according to perceived
values. Ideally, the level of false accept risk is established to minimize total cost. In developing a false accept baseline, the benefit to be derived from maintaining a given level of false accept risk should be weighed against the cost of achieving it.
There are two ways to express false accept risk:
Consumer's False Accept Risk
Consumer's false accept risk is the probability that an out-of-tolerance parameter will be found in the population of accepted parameters. One way of looking at consumer's false accept risk is to imagine it to be the number of out-of-tolerance parameters delivered to the user (consumer) divided by the total number of parameters delivered.
Consumer's false accept risk is a useful statistic for evaluating "out the door" measurement quality.
This option should be used if the concern is evaluating performance from the customer's or user's perspective.
Producer's False Accept Risk (Option 2)
Producer's false accept risk is the probability that a measured out-of-tolerance parameter will be called in-tolerance during measurement. This risk can be envisioned as the number of out-of-tolerance parameters delivered divided by the total number of parameters measured.
false accept risk is a useful statistic for evaluating how well the measuring organization (the producer) is performing.
This option can also be used to evaluate performance from the customer's or
user's perspective under certain conditions.
Note: When you change the false accept risk option and click OK, the baseline and computed false accept risks are automatically recalculated.
False reject risk is manifested in terms of costs associated with rejected items. These costs may appear as repair, rework, re-test, re-calibration, etc. costs. The impact of false reject risk is fairly straightforward to evaluate. This impact is expressed as the total number of items (parameters) tested multiplied by the cost of a rejection multiplied by the false reject risk.
Establishing a false reject risk baseline can often be a simple matter of weighing operating costs against operating budgets.
Show Defaults on the menu displays baseline false accept and false
reject risks for a 4:1 accuracy ratio with 95% in-tolerance probabilities
for the subject parameter and measuring system. Clicking the Use
Default Baseline button automatically applies the default baseline
risks, accuracy ratio and in-tolerance probabilities to the corresponding
Analysis Baseline values. Conversely, clicking the Set Default
Baseline button automatically applies the Analysis Baseline accuracy
ratio, in-tolerance probabilities and risks to the corresponding Default
Risk vs. Baseline Accuracy Ratio Plot
A plot of the analysis baseline false accept risk and false reject risk vs. accuracy ratio may be generated by clicking
View Plot on the menu. This activates the Risk vs. Baseline Accuracy Ratio
Screen. The Risk vs. Baseline Accuracy Ratio plot may be copied to the Windows clipboard or printed.
If desired, tabulated data from the plot can also be copied to the Windows