Causes of Variation in Quality

No two products or service outputs are ever exactly identical; some degree of variation is always present. Understanding the sources of this variation is fundamental to quality control, as the goal is often to reduce undesirable variation. Variation typically arises from two main types of causes:

1. Chance Causes (Common Causes / Random Variation):

• Definition: These are the many small, inherent, unavoidable sources of variation that exist within any stable process. They are numerous, individually insignificant, and difficult or uneconomical to eliminate completely.
• Characteristics:
  • Affect all outputs of the process.
  • Result from the inherent design and operation of the system (e.g., slight fluctuations in machine vibration, temperature, humidity, material consistency, operator reaction time).
  • Produce a predictable, stable pattern of variation over time, often following a statistical distribution (like the normal distribution).
  • Cannot be traced to a single specific cause.
• Control Implication: When only chance causes are present, the process is considered "in statistical control." Reducing this type of variation typically requires fundamental changes to the process itself (e.g., better equipment, improved design, tighter material specs), which is usually a management responsibility.
• Control Chart Appearance: Points fluctuate randomly within the control limits.

2. Assignable Causes (Special Causes / Non-Random Variation):

• Definition: These are specific, identifiable factors that cause larger, unpredictable variations outside the normal, stable pattern expected from the process.
• Characteristics:
  • Not always present in the process; they arise intermittently or due to specific events.
  • Can often be traced to a particular source (e.g., operator error, incorrect machine setting, a bad batch of material, a worn tool, power failure, malfunctioning sensor).
  • Cause the process output to shift or become unstable.
  • Can and should be identified and eliminated.
• Control Implication: When assignable causes are present, the process is considered "out of statistical control." The goal of SPC is to detect the presence of assignable causes quickly so that they can be investigated and corrected, bringing the process back into a state of statistical control. This is often the responsibility of process operators and supervisors.
• Control Chart Appearance: Points fall outside control limits, or exhibit non-random patterns (runs, trends, cycles) within the limits.

Indian Example: In a chai stall in India making tea: * Chance Causes of variation in taste might include minor differences in tea leaf strength from the same batch, slight variations in boiling time, tiny fluctuations in the amount of milk or sugar added even with careful pouring. The tea is generally consistent but never exactly identical. * Assignable Causes could be: the usual person making tea is absent and a new person makes it differently (Manpower), the gas runs low affecting heat (Machine/Energy), a new brand of milk with different fat content is used (Material), the sugar container was accidentally filled with salt (Error/Method). These cause a noticeable, significant change in the tea's quality and need to be identified and fixed.

Distinguishing between these two types of variation using tools like control charts is the key to effective Statistical Process Control. We address assignable causes to achieve stability, and then work on reducing common cause variation to improve overall process capability.