Shewhart Control Chart Basics
You'll learn how Shewhart control charts use three simple lines to distinguish normal process variation from genuine anomalies.
The Three-Line Structure
A Shewhart control chart (also called a process control chart) plots your data points over time and overlays three horizontal reference lines that define normal behavior:
- Center Line (CL): Represents the process mean or target value—where your process should be centered when operating normally
- Upper Control Limit (UCL): The upper boundary of expected variation
- Lower Control Limit (LCL): The lower boundary of expected variation
Think of these like highway lane markers. The center line is your ideal lane position, while the control limits are the lane edges. Drifting between the lines? That's normal driving variation. Cross outside the lines? That signals something unusual requiring attention.
How the Limits Work
The control limits are typically set at ±3 standard deviations from the center line. This captures approximately 99.7% of normal process variation (assuming a normal distribution). Points falling outside these limits are flagged as potential anomalies—they're statistically unlikely to occur by random chance alone.
The beauty of this approach is its visual simplicity: you can immediately see when a process shifts out of control, rather than calculating statistics for every data point.
The Foundation for Monitoring
Unlike static outlier detection methods you've learned (Z-score, IQR, Grubbs' test), Shewhart charts are designed for ongoing monitoring of processes over time. They don't just ask "is this point unusual?" but "is this process behaving consistently?"
Key Takeaway: Shewhart control charts use three reference lines—a center line for the process mean and upper/lower control limits (typically ±3σ)—to visually separate normal variation from statistically significant anomalies in time-ordered data.