CQI key terms

Sum of all values in a data set divided by the number of values; a measure of central tendency used to estimate the process average.

Middle value in a data set arranged in order; if the count is even, the average of the two middle values. Less affected by extreme outliers than the mean.

Measure of the spread of values around the mean, computed as the square root of the variance. Small values indicate tightly clustered data.

Average of the squared differences between each value and the mean; the square of the standard deviation. Used in capability and control chart calculations.

Symmetric, bell-shaped probability distribution defined by its mean and standard deviation; about 68% of values fall within one standard deviation of the mean.

Difference between the largest and smallest values in a data set. Simple measure of spread often plotted on control charts for subgroups.

Ratio comparing the width of the specification limits to the natural spread of a stable process; Cpk also accounts for how well the process is centered.

Value below which a given percentage of observations in a data set falls; the 50th percentile equals the median. Used to describe distributions and set limits.

Quantified range around a measured value within which the true value is expected to lie, combining contributions from the instrument, environment, operator and procedure.

Comparison of an instrument's readings against a reference standard of known value, documenting the deviation and adjusting the instrument when necessary.

Unbroken chain of comparisons linking a measurement result to national or international standards, each step documented with a stated uncertainty.

Smallest change in a quantity that an instrument can display or detect. Finer resolution does not by itself guarantee better accuracy.

Closeness of a measured value to the true or accepted reference value. Accuracy describes the overall correctness of a reading, not its consistency.

Closeness of repeated measurements to one another under unchanged conditions. A precise instrument can still be inaccurate if it is biased.

Variation in measurements taken by one operator using the same instrument on the same item under the same conditions, often called equipment variation.

Variation in measurements caused by different operators, setups or laboratories using the same instrument and procedure; often called appraiser variation.

Consistent difference between the average of repeated measurements and a reference value; a systematic error that shifts all readings in the same direction.

Change in an instrument's bias across its operating range; a linear instrument has the same error at low and high readings.

Study that estimates how much measurement variation comes from the gauge (repeatability) and from operators (reproducibility), usually expressed as a percentage of tolerance or process variation.

Device or material with a known, stable value traceable to a higher authority, used to calibrate or verify other measuring instruments.

Deciding whether to accept a lot by inspecting a sample rather than every unit, using a plan that balances producer and consumer risk.

Documented rule stating sample size, acceptance number and rejection number for deciding on a lot; may be based on attributes or variables.

The quality level that is the worst tolerable process average when a continuing series of lots is submitted for acceptance sampling.

Checking each unit only for conformance or nonconformance to a requirement, such as go/no-go gauging or visual defect counting, rather than measuring a numeric value.

Measuring a numerical characteristic of a unit and comparing it to specification limits, producing data that can be analyzed statistically.

Full verification of the first production unit against all drawing and specification requirements before the remainder of the lot is manufactured.

Checks performed during manufacturing at defined stages to detect and contain nonconformities before further value is added.

Last verification of finished product against all requirements, performed after all production steps are complete and before release or shipment.

Product that does not meet one or more requirements; it must be identified, segregated and dispositioned as rework, repair, scrap, or use-as-is with approval.

Attribute gauge with two members; the go member must fit and the no-go member must not, verifying that a feature lies within its limits.

Planned activities and system-level processes intended to prevent defects and give confidence that requirements will be met, rather than detecting defects after the fact.

Operational techniques and inspections used to verify that products meet requirements and to stop nonconforming output from moving forward.

Action taken to eliminate the cause of a detected nonconformity so it does not recur, usually following root cause analysis.

Action taken to eliminate the cause of a potential nonconformity before it occurs, based on analysis of trends and risk data.

Structured investigation to determine the underlying reason a problem occurred, distinguishing it from symptoms so that corrective action addresses the real cause.

Bar chart ordering categories by frequency with a cumulative line, used to focus improvement effort on the few causes responsible for most problems.

Diagram organizing possible causes of a problem into categories such as people, methods, machines, materials, measurement and environment, often called a fishbone.

Graph of a process measure over time with statistically calculated control limits, used to distinguish common-cause variation from special-cause signals.

Systematic, independent examination of processes, records or products against defined criteria to verify that the quality system is implemented and effective.

Identification on product or records showing which inspections and tests have been passed, failed or are still pending, preventing premature release.