Course contentsShow
← Data Science
Lesson 1402 of 2,14535. Statistical Anomaly DetectionFree lesson

Western Electric Rules

Apply additional run rules to detect non-random patterns in control charts before major shifts occur.

Western Electric Rules

You'll learn how to catch subtle warning signs in process data before they turn into major problems.

What Are Western Electric Rules?

The Western Electric Rules are a set of pattern-recognition guidelines developed in the 1950s to detect non-random behavior in control charts. While the basic Shewhart chart (which you already know) flags points outside control limits, these rules catch early warning patterns that suggest something is drifting before crossing those hard boundaries.

Think of it like monitoring your health: waiting until you have a fever (out-of-control signal) is reactive. Noticing you've felt "slightly off" for three days straight (a Western Electric pattern) lets you act earlier.

The Four Core Rules

Rule 1: One point beyond 3σ (you already know this from Shewhart charts)

Rule 2: Two out of three consecutive points beyond 2σ on the same side of the centerline

Rule 3: Four out of five consecutive points beyond 1σ on the same side

Rule 4: Eight consecutive points on one side of the centerline (even if within 1σ)

These rules detect trends, shifts, and clustering that are statistically unlikely under normal variation. Rule 4, for example, would catch a gradual upward drift that hasn't yet breached the upper control limit.

Why They Matter

Random processes should bounce around the centerline unpredictably. When you see systematic patterns—too many consecutive points on one side, clustering near control limits—it suggests an assignable cause has entered your process, even if individual points haven't violated basic limits yet.

Key Takeaway: Western Electric Rules extend basic control chart detection by identifying suspicious patterns (runs, trends, clustering) that signal process shifts before they become severe violations.