Course contentsShow
← Data Science
Lesson 1401 of 2,14535. Statistical Anomaly DetectionFree lesson

Detecting Out-of-Control Signals

Recognize patterns indicating special cause variation: points beyond limits, runs, and trend signals.

Detecting Out-of-Control Signals

What you'll learn: How to recognize specific patterns in control charts that indicate your process has shifted from normal variation to special cause variation requiring investigation.

The Core Idea

You've learned how to build control charts with control limits, but simply drawing the chart isn't enough. You need to know what patterns to look for that scream "something unusual is happening!" These patterns are called "out-of-control signals," and they indicate special cause variation—shifts, trends, or anomalies that aren't part of your process's normal random fluctuation.

Think of it like monitoring your car's dashboard. It's not just about whether the temperature gauge crosses into the red zone—you'd also be concerned if it steadily climbed over time, or if it bounced unusually high and low repeatedly.

Key Patterns to Watch For

1. Points Beyond Control Limits

The most obvious signal: any single point outside the upper or lower control limit indicates the process is out of control.

2. Runs

A sequence of consecutive points all on one side of the centerline (typically 7+ points). This suggests the process mean has shifted, even if no points crossed the limits.

3. Trend Signals

A steady upward or downward pattern (typically 7+ points consistently increasing or decreasing). This indicates a drift in the process over time.

4. Other Western Electric Rules

Two out of three consecutive points in the outer third of the chart, or four out of five in the middle-to-outer region also signal non-random patterns.

Key Takeaway: Out-of-control signals are specific, recognizable patterns—points beyond limits, runs on one side of center, and sustained trends—that distinguish special cause variation from normal process randomness, triggering investigation and corrective action.