Control charts are not perfect tools for detecting shifts in the process distribution because they are

based on sampling distributions. Two types of error are possible with the use of control charts. A
type I error occurs when the conclusion is made that the process is out of control based on a
sample result that falls outside the control limits, when in fact it was due to pure randomness. A
type II error occurs when the conclusion is that the process is in control and only randomness
is present, when actually the process is out of statistical control.

Control Charts for Variables

Control charts for variables are used to monitor the mean and the variability of the
process distribution

R-Chart A range chart, or R-chart, is used to monitor process variability. To calculate

the range of a set of sample data, the analyst subtracts the smallest from the largest
measurement in each sample. If any of the ranges fall outside the control limits, the
process variability is not in control. The control limits for the R-chart are

Notice that the values for D3 and D4 shown in Table 3.1 change as a function of the
sample size.
Notice, too, that the spread between the control limits narrows as the sample size
increases. This change is a consequence of having more information on which to base an
estimate for the process range.
x-Chart An x-Chart (read “x-bar chart”) is used to see whether the process is
generating output, on average, consistent with a target value set by management for the
process or whether its current performance, with respect to the average of the
performance measure, is consistent with its past performance. A target value is useful
when a process is completely redesigned and past performance is no longer relevant.
When the assignable causes of process variability have been identified and the process
variability is in statistical control, the analyst can then construct an x-chart. The control
limits for the x-chart are

Step 1. Collect data on the variable quality measurement (such as time, weight, or
diameter) and organize the data by sample number. Preferably, at least 20 samples of size
n should betaken for use in constructing a control chart.
Step 2. Compute the range for each sample and the average range, R, for the set of
samples.
Step 3. Use Table 3.1 to determine the upper and lower control limits of the R-chart.
Step 4. Plot the sample ranges. If all are in control, proceed to step 5. Otherwise, find
the assignable causes, correct them, and return to step 1.
Step 5. Calculate x for each sample and determine the central line of the chart, x.
Step 6. Use Table 3.1 to determine the parameters for UCLx and LCLx and construct the
x-chart.
Step 7. Plot the sample means. If all are in control, the process is in statistical control in
terms of the process average and process variability. Continue to take samples and
monitor the process. If any are out of control, find the assign-able causes, address them,
and return to step 1. If no assignable causes are found after a diligent search, assume that
the out-of-control points represent common causes of variation and continue to monitor
the process.
Using x and R-Charts to Monitor a Process

The management of West Allis Industries is concerned about the production of a special
metal screw used by several of the company’s largest customers. The diameter of the
screw is critical to the customers. Data from five samples appear in the accompanying
table. The sample size is 4. Is the process in statistical control?
DECISION POINT
A new employee operated the lathe machine that makes the screw on the day sample 5
was taken. To solve the problem, management initiated a training session for the
employee. Subsequent samples showed that the process was back in statistical control.

Control Charts for Attributes

Two charts commonly used for performance measures based on attributes measures are
the p- and c-chart. The p-chart is used for controlling the proportion of defects generated
by the process. The c-chart is used for controlling the number of defects when more than
one defect can be present in a service or product.
p-Charts The p-chart is a commonly used control chart for attributes. The performance
characteristic is counted rather than measured, and the entire service or item can be
declared good or defective. For example, in the banking industry, the attributes counted
might be the number of non-endorsed deposits or the number of incorrect financial
statements sent to customers. The method involves selecting a random sample, inspecting
each item in it, and calculating the sample proportion defective, p, which is the number
of defective units divided by the sample size.
Sampling for a p-chart involves a “yes or no” decision: The process output either is or is
not defective. The underlying statistical distribution is based on the binomial distribution.
However, for large sample sizes, the normal distribution provides a good approximation
to it. The standard deviation of the distribution of proportion defectives,

The chart is used in the following way. Periodically, a random sample of size n is taken,
and the number of defective services or products is counted. The number of defectives is
divided by the sample size to get a sample proportion defective, p, which is plotted on the
chart. When a sample proportion defective falls outside the control limits, the analyst
assumes that the proportion defective generated by the process has changed and searches
for the assignable cause. Observations falling below the LCLp indicate that the process
may actually have improved. The analyst may find no assignable cause because it is
always possible that an out-of-control proportion occurred randomly. However, if the
analyst discovers assignable causes, those sample data should not be used to calculate the
control limits for the chart.

Using a p-Chart to Monitor a Process

The operations manager of the booking services department of Hometown Bank is
concerned about the number of wrong customer account numbers recorded by Hometown
personnel. Each week a random sample of 2,500 deposits is taken, and the number of
incorrect account numbers is recorded. The results for the past 12 weeks are shown in the
following table. Is the booking process out of statistical control? Use three-sigma control
limits, which will provide a type I error of 0.26 percent.
Step 3. Plot each sample proportion defective on the chart, as shown in Figure 3.11

Sample 7 exceeds the UCL; thus, the process is out of control and the reasons for the
poor performance that week should be determined.

DECISION POINT
Management explored the circumstances when sample 7 was taken. The encoding
machine used to print the account numbers on the checks was defective that week. The
following week the machine was repaired; however, the recommended preventive
maintenance on the machine was not performed for months prior to the failure.
Management reviewed the performance of the maintenance department and instituted
changes to the maintenance procedures for the encoding machine. After the problem was
corrected, an analyst recalculated the control limits using the data without sample 7.
Subsequent weeks were sampled, and the booking process was determined to be in
statistical control. Consequently, the p-chart provides a tool to indicate when a process
needs adjustment.

c-Charts Sometimes services or products have more than one defect. For example, a roll
of carpeting may have several defects, such as tufted or discolored fibers or stains from
the production process. Other situations in which more than one defect may occur
include accidents at a particular intersection, bubbles in a television picture face panel,
and complaints from a patron at a hotel. When management is interested in reducing the
number of defects per unit or service encounter, another type of control chart, the c-chart,
is useful. The underlying sampling distribution for a c-chart is the Poisson distribution.
The Poisson distribution is based on the assumption that defects occur over a continuous
region on the surface of a product or a continuous time interval during the provision of a
service. It further assumes that the probability of two or more defects at any one location
on the surface or at any instant of time is negligible. The mean of the distribution is c and
the standard deviation is 2c. A useful tactic is to use the normal approximation to the
Poisson so that the central line of the chart is c and the control limits are

Using a c-Chart to Monitor Defects per Unit

The Woodland Paper Company produces paper for the newspaper industry. As a final
step in the process, the paper passes through a machine that measures various product
quality characteristics. When the paper production process is in control, it averages 20
defects per roll.
a. Set up a control chart for the number of defects per roll. For this example, use two-
sigma control limits.
b. Five rolls had the following number of defects: 16, 21, 17, 22, and 24, respectively.
The sixth roll, using pulp from a different supplier, had 5 defects. Is the paper production
process in control?

a. The average number of defects per roll is 20. Therefore,

b. Because the first five rolls had defects that fell within the control limits, the process is
still in control. The sixth roll’s five defects, however, is below than the LCL, and
therefore, the process is technically “out of control.” The control chart indicates that
something good has happened.

DECISION POINT

The supplier for the first five samples has been used by Woodland Paper for many years.
The supplier for the sixth sample is new to the company. Management decided to
continue using the new supplier for a while, monitoring the number of defects to see
whether it stays low. If the number remains below the LCL for 20 consecutive samples,
management will make the switch permanent and recalculate the control chart
parameters.