I-MR / X-MR Chart
An I-MR chart monitors individual measurements and their moving ranges. It is the correct Shewhart chart when each time point has one observation rather than a rational subgroup.[1]
Prerequisites
Prerequisites: control chart basics and ordered observations.
Process Context
Use I-MR for low-volume production, destructive tests, long cycle times, administrative processes, or any setting where subgrouping is not meaningful.
Definition
The individuals chart plots
Assumptions / Requirements
- One observation per time point.
- Consecutive observations are meaningful for estimating short-term variation.
- The process is not strongly autocorrelated.
- Measurement resolution is adequate.
- Phase I data are screened for special causes before final limits are used.
Notation
| Symbol | Meaning |
|---|---|
|
|
Individual observation at time
|
|
|
Average of individual observations |
|
|
Moving range between
|
|
|
Average moving range |
|
|
Range constant,
|
Control Limits / Formula
For the individuals chart:
For the moving range chart with range length 2:
Interpretation Rules
- Check the moving range chart for unstable short-term variation.
- Investigate individual points outside limits and nonrandom patterns.
- If autocorrelation is visible, use a time-series or residual approach instead of naive I-MR limits.
- Never use specification limits like chart limits.
Worked Example
Five individual measurements are
No listed point or moving range exceeds these limits.
Common Mistakes
- Calling an individuals chart an Xbar chart.
- Creating artificial subgroups after the fact.
- Ignoring autocorrelation from automated or continuous processes.
- Using a moving range when consecutive observations are not comparable.
Connections
| Related note | Use |
|---|---|
| Control charts | Chart selection |
| Xbar-R chart | True subgroup chart |
| Autocorrelated data | When I-MR assumptions fail |
| Control Limits and Specification Limits | Avoid limit confusion |
References
NIST/SEMATECH, e-Handbook of Statistical Methods, "Individuals Control Charts", https://www.itl.nist.gov/div898/handbook/pmc/section3/pmc322.htm ↩︎