P chart

P chart
See also

P chart, also known as a fraction nonconforming control chart, is a graphical tool which was developed in industry to help with interpreting and reducing sources of variability in manufacturing processes. The definition of fraction nonconforming (defective) (p) is usually exemplified by the ratio of the number of nonconforming items in the population to the total number of items in that population. Investigated item may have one or more quality characteristics that are examined simultaneously. If at least one of the attributes does not conform to standard, the item is classified as nonconforming [1].


In the 1920s, W.A. Shewhart developed the concept of statistical control chart for the purpose of improving the reliability of telephone transmission systems. The conception arose from observating the operators. Shewhart noticed that they were overreacting and making inappropriate changes in settings as a response to indicator variations that were simply random. Those decisions were wasteful and inducted more variation in the process, making the system less stable. Shewhart's theory of variation provides that quality is inversely proportional to variability. It also states that understanding the variability of some indicators may show the operator when and how reduce it [2].

At that time, the concept of using the two-point moving range for measuring the dispersion of a set of individual values didn't occurred yet. Shewhart faced the problem how to create a process behavior chart for individual values based on counts. Then he decided to use theoretical limits based on a probability model. Therefore, he could estimate both the central line and the three-sigma distance with only one location statistic[3].

Design of the P chart

There are three parameters of fraction nonconforming control chart that must be specified: the sample size, the frequency of sampling, and the width of the control limits. Preferably, we should have some general guidance for selecting those parameters.

Formulas for the points on the Chart

Let's suppose we have k samples, each of ni size. Let Di stand for the number of nonconforming units in the ith sample. The ith proportion pi is calculated as the following ratio[4]\[p_{i}=\frac{D_{i}}{n_{i}}\].

P Chart Center Line

In the P Charts procedure, the center line proportion may be implement directly. It can also be estimated from a series of samples. If it is estimated from the samples we use the formula for the centerline proportion\[\bar{p}=\frac{\sum_{i=1}^{k}D_{i}}{\sum_{i=1}^{k}n_{i}}\].

You can reduce this formula if all the samples are the same size[5].\[\bar{p}=\frac{\sum_{i=1}^{k}D_{i}}{kn}=\frac{\sum_{i=1}^{k}p_{i}}{k}\]

P Chart limits

To calculate the lower and upper control limits for the P chart use these formulas\[LCL=\bar{p}-m\sqrt{\frac{\bar{p}(1-\bar{p})}{n_{i}}}\]\[UCL=\bar{p}+m\sqrt{\frac{\bar{p}(1-\bar{p})}{n_{i}}}\]

where m stands for multiplier (usually set to 3) chosen to control the probability of false alarms (out-of-control signals when the process is under control)[6].

There are two steps that one should take into concideration to avoid potential pitfalls[7]:

  • Ensure that there are enough observations taken for each sample,
  • Be wary of differences in the number of observations from each sample.


  1. Nist/Sematech…, 2012
  2. Duclos A., Voirin N., 2010, p. 402
  3. Wheeler D. J., 2011
  4. Montgomery D.C., 2012, pp. 298-309
  5. Montgomery D.C., 2012, pp. 298-309
  6. Montgomery D.C., 2012, pp. 298-309
  7. Ryan T.P., 2011, 184-188


Author: Anna Kasprzyk