6.3: Process stability and predictable performance

Quality and Regulatory Compliance in Bioscience · LibreTexts

0%

Contents

    Text

    Size
    100%
    Theme
    Typeface
    Line length
    Spacing

    Read aloud

    How

    Speed: 1×
    Light up
    Highlight

    Symbols: Mulberry Symbols, © 2018–2026 Steve Lee, CC BY-SA 4.0 (licence, the set), the site's own copies.

    From LibreTexts, under CC BY-SA 4.0, as the source states it; its equations are written in TeX, which the PDF shows typeset. This page at LibreTexts.

    6.3: Process stability and predictable performance

    A stable process has a reasonably consistent pattern of results over time under the conditions studied. Statistical control means that the pattern is consistent with common-cause variation. NIST describes stability through a consistent average, variability, and distribution over time. This supports predictions under similar conditions, while leaving room for future changes.Process Stability and Control Charts.png

    The mean is the arithmetic average: add the values and divide by their number. Spread describes how widely values differ. Two processes can have the same mean but very different spreads. A filler alternating between low and high volumes may average 10.00 mL while producing bottles that are individually unsuitable.

    Stability supports conditional prediction. If relevant conditions remain the same, the established pattern provides a basis for estimating future behavior. It does not guarantee the next exact value or rule out a future failure. A stable process also can perform poorly in a consistent way.

    A control chart plots a process measurement or summary in time order and compares it with a reference pattern. The chart includes a center line and appropriate control limits derived using process data and a suitable statistical method. Some charts monitor individual values; others monitor subgroup averages, ranges, or proportions. Select a chart that matches the data and sampling design.

    A team looks for defined signals: patterns that warrant investigation under the chosen chart rules. Examples include a point beyond a control limit or a sustained pattern on one side of the center line. A signal can occur while every point remains inside the limits. Establish suitable rules in advance. A signal is evidence to investigate, rather than proof of a particular cause, and a chart can occasionally signal when no process change has occurred. 

    A short sequence that looks quiet does not establish long-term stability. Baseline data should represent relevant shifts, material lots, and routine operating conditions for the defined process. Keep unlike processes separate. The sampling frequency and relationships between successive measurements also affect interpretation. An apparent pattern may reflect how samples were collected rather than a change in production.

    An unstable process shows statistical evidence that its monitored behavior is changing. Combining changing conditions into one average can hide important information. Investigate before using pooled data to predict future performance. A chart supports judgment; technical knowledge, reliable measurement, and a quality assessment remain necessary.


    This page titled 6.3: Process stability and predictable performance was last modified on Fri, 06 Nov 2026 23:31:42 GMT and is shared under a CC BY-SA 4.0 license and was authored, remixed, and/or curated by Emalee MacKenzie.