Step 1
What are you measuring?
Step 2 · Subgroup size
Step 2 · What are you counting?
Step 3 · Sample consistency
Calculator Library / SPC
Control Chart Selector
Answer two or three quick questions about your data and get the right control chart back, with its exact formulas — no memorizing the decision tree required.
Decision Map
How the answer was reached
Instructions
How to use this app
- Start with what you're measuring: a continuous number on a scale, or an attribute count.
- Answer the follow-up question — subgroup size for continuous data, or defective-unit vs. defect-count for attribute data.
- For attribute data, answer whether the sample size or inspection unit stays constant or varies from subgroup to subgroup.
- Read the recommended chart, its formulas, and follow the link to build real control limits from your own data.
- Use the quick-scenario chips to see how a few common shop-floor situations map to a chart.
What This Control Chart Selector Does
Choosing the wrong control chart is one of the most common SPC mistakes, and it's rarely obvious after the fact — the chart still draws lines, they just aren't the right lines. This tool walks through the same two or three questions an experienced quality engineer asks before picking a chart: what kind of data is this, how is it sampled, and does that sampling stay consistent.
It's built to pair with the Control Limits Generator, which calculates the actual center line and control limits once the right chart type is known.
Core Decision Logic
| Data type | Sampling detail | Chart |
|---|---|---|
| Continuous | One reading at a time (n=1) | I-MR |
| Continuous | Small subgroups (n=2–10) | X-bar / R |
| Continuous | Large subgroups (n>10) | X-bar / S |
| Attribute — defective units | Constant sample size | np |
| Attribute — defective units | Varying sample size | p |
| Attribute — defect counts | Constant inspection unit | c |
| Attribute — defect counts | Varying inspection unit | u |
Worked Examples
Ridgeline Precision Machining, the shop used throughout the SPC & Control Charts Guide, uses two different charts on the same production line.
Their CNC cell pulls a subgroup of 5 bore-diameter measurements every hour — continuous data with small, consistent subgroups, which this tool correctly routes to an X-bar/R chart. Their final visual inspection station instead checks a fixed sample of 100 finished housings each day and simply marks each one pass or fail — attribute data, defective units, constant sample size, which routes to an np chart instead.
How to Interpret the Results
- I-MR: the fallback when there's no natural subgroup — slower to detect small shifts than a subgrouped chart, but usable with a single reading.
- X-bar/R vs. X-bar/S: both track the same two things, mean and spread; the crossover at n≈10 is about estimator efficiency, not a hard rule.
- np vs. p, c vs. u: in each pair, the second chart exists purely to handle a sample size that isn't fixed.
- When in doubt: a chart built on the wrong assumption produces control limits that are either too tight (false alarms) or too loose (missed signals) — getting this choice right matters before any data collection starts.
Control Chart Selector Frequently Asked Questions
What's the difference between counting defective units and counting defects?
A defective unit is a pass/fail judgment on the whole unit, even if it failed for more than one reason. A defect count tallies every individual nonconformance, so one unit can carry several defects. Pass/fail data points toward p or np charts; defect-tally data points toward c or u charts.
Why does subgroup size decide between the range and the standard deviation?
The range is fast to compute by hand and works well for small subgroups, but it throws away information as subgroups get larger, since it only looks at the highest and lowest value. Past roughly n=10, the standard deviation uses every point in the subgroup and becomes the more statistically efficient estimator, which is why X-bar/S takes over from X-bar/R at that point.
Can a process switch chart types partway through a project?
Yes, and it's common. A process often starts on an I-MR chart with individual readings while a team is still small-scale checking, then moves to X-bar/R once subgroup sampling is set up. Restart the control limits calculation from fresh data whenever the chart type changes, since limits from one chart type don't carry over to another.
What if the sample size or inspection unit size isn't fixed in advance?
That's exactly the signal to use the varying-sample-size chart in each pair: a p chart instead of np for defective units, or a u chart instead of c for defect counts. Both recalculate their control limits for each subgroup's own sample size, so they stay valid even when the count isn't constant.
Related Templates and Guides
Read the SPC & Control Charts Guide
The full theory behind every chart on this page, plus Western Electric rules and a hand-worked X-bar/R example.
Open the Control Limits Generator
Once you know your chart type, build the actual center line and control limits from your own subgroup data.